Clash Royale CLAN TAG#URR8PPP

up vote
23
down vote

favorite

Given an N-dimensional orthogonal (non-ragged) array of non-negative integers, and an indication of which dimensions to reverse, return the array but reversed along those dimensions. The indication may be given as a Boolean list of length N or a list of a subset of the first N dimensions indexed from 0 or 1.

Please state your input formats. Code explanations are much appreciated.

Walked-through example

We are given the 2-layer 3-row 4-column 3D-array

[[[ 1, 2, 3, 4],
 [ 5, 6, 7, 8],
 [ 9,10,11,12]],

 [[13,14,15,16],
 [17,18,19,20],
 [21,22,23,24]]]

and one of

[true,false,true] (Boolean list)
[0,2] (0-indexed list)
[1,3] (1-indexed list)

We need to reverse order of the first and last dimensions, that is the layers and the elements of the rows (the columns), but not the rows of each layer. First (the actual order you do this in does not matter) we reverse the order of the layers:

[[[13,14,15,16],
 [17,18,19,20],
 [21,22,23,24]],

 [[ 1, 2, 3, 4],
 [ 5, 6, 7, 8],
 [ 9,10,11,12]]]

and then we reverse the order of the elements of each row:

[[[16,15,14,13],
 [20,19,18,17],
 [24,23,22,21]],

 [[ 4, 3, 2, 1],
 [ 8, 7, 6, 5],
 [12,11,10, 9]]]

Test cases

[[[1,2,3,4],[5,6,7,8],[9,10,11,12]],[[13,14,15,16],[17,18,19,20],[21,22,23,24]]]
[true,false,true]/[0,2]/[1,3]

Ã¢Â€ÂƒÃ¢Â†Â“Ã¢Â€Âƒ
[[[16,15,14,13],[20,19,18,17],[24,23,22,21]],[[4,3,2,1],[8,7,6,5],[12,11,10,9]]]

[[1,2,3],[4,5,6]]
[true,false]/[0]/[1]

Ã¢Â€ÂƒÃ¢Â†Â“
[[4,5,6],[1,2,3]]

[[1],[4]]
[true,false]/[0]/[1]

Ã¢Â€ÂƒÃ¢Â†Â“
[[4],[1]]

[[7]]
[true,true]/[0,1]/[1,2]

Ã¢Â€ÂƒÃ¢Â†Â“
[[7]]

[1,2,3,4,5,6,7]
[true]/[0]/[1]

Ã¢Â€ÂƒÃ¢Â†Â“
[7,6,5,4,3,2,1]

[true]/[0]/[1]

Ã¢Â€ÂƒÃ¢Â†Â“

[,]
[false,false]//

Ã¢Â€ÂƒÃ¢Â†Â“
[,]

[[[[3,1,4,1],[5,9,2,6]],[[5,3,5,8],[9,7,9,3]]],[[[2,3,8,4],[6,2,6,4]],[[3,3,8,3],[2,7,9,5]]]]
[true,false,true,true]/[0,2,3]/[1,3,4]

Ã¢Â€ÂƒÃ¢Â†Â“
[[[[4,6,2,6],[4,8,3,2]],[[5,9,7,2],[3,8,3,3]]],[[[6,2,9,5],[1,4,1,3]],[[3,9,7,9],[8,5,3,5]]]]

[[[[3,1,4,1],[5,9,2,6]],[[5,3,5,8],[9,7,9,3]]],[[[2,3,8,4],[6,2,6,4]],[[3,3,8,3],[2,7,9,5]]]]
[false,true,false,false]/[1]/[2]

Ã¢Â€ÂƒÃ¢Â†Â“
[[[[5,3,5,8],[9,7,9,3]],[[3,1,4,1],[5,9,2,6]]],[[[3,3,8,3],[2,7,9,5]],[[2,3,8,4],[6,2,6,4]]]]

[[[[3,1,4,1],[5,9,2,6]],[[5,3,5,8],[9,7,9,3]]],[[[2,3,8,4],[6,2,6,4]],[[3,3,8,3],[2,7,9,5]]]]
[false,false,false,false]//

Ã¢Â€ÂƒÃ¢Â†Â“
[[[[3,1,4,1],[5,9,2,6]],[[5,3,5,8],[9,7,9,3]]],[[[2,3,8,4],[6,2,6,4]],[[3,3,8,3],[2,7,9,5]]]]

edited Sep 3 at 13:54

asked Sep 3 at 12:44

AdÃ¡m

27.7k268185

I feel like the hardest part in most statically typed languages will be golfing the type signatures involved.
â€“Â ÃŽÂŸurous
Sep 3 at 22:07

@ÃŽÂŸurous How do those languages normally deal with arbitrary array data?
â€“Â AdÃ¡m
Sep 3 at 22:29

1

there's three cases for "normal" use as I see it: only worrying about one level of the array (eg: reverse works on arbitrary arrays but only cares about the first level), generics, or recursive classes (type / object classes depending on functional or OOP, but similar use-case). The latter two are usually far more verbose.
â€“Â ÃŽÂŸurous
Sep 3 at 22:39

Can we store the matrix as arrays of pointers to pointers (in C or asm), instead of proper multi-dimensional arrays where everything is contiguous in memory? I'm pretty sure all the normal higher-level / dynamically-typed languages with arbitrary nesting of lists are already treating things as lists of lists, not matrices, so I'm going to assume it's fine.
â€“Â Peter Cordes
Sep 5 at 3:42

@PeterCordes Sure, go ahead.
â€“Â AdÃ¡m
Sep 5 at 5:35

add a commentÂ |Â

up vote
23
down vote

favorite

Please state your input formats. Code explanations are much appreciated.

Walked-through example

We are given the 2-layer 3-row 4-column 3D-array

[[[ 1, 2, 3, 4],
 [ 5, 6, 7, 8],
 [ 9,10,11,12]],

 [[13,14,15,16],
 [17,18,19,20],
 [21,22,23,24]]]

and one of

[true,false,true] (Boolean list)
[0,2] (0-indexed list)
[1,3] (1-indexed list)

[[[13,14,15,16],
 [17,18,19,20],
 [21,22,23,24]],

 [[ 1, 2, 3, 4],
 [ 5, 6, 7, 8],
 [ 9,10,11,12]]]

and then we reverse the order of the elements of each row:

[[[16,15,14,13],
 [20,19,18,17],
 [24,23,22,21]],

 [[ 4, 3, 2, 1],
 [ 8, 7, 6, 5],
 [12,11,10, 9]]]

Test cases

[[1,2,3],[4,5,6]]
[true,false]/[0]/[1]

Ã¢Â€ÂƒÃ¢Â†Â“
[[4,5,6],[1,2,3]]

[[1],[4]]
[true,false]/[0]/[1]

Ã¢Â€ÂƒÃ¢Â†Â“
[[4],[1]]

[[7]]
[true,true]/[0,1]/[1,2]

Ã¢Â€ÂƒÃ¢Â†Â“
[[7]]

[1,2,3,4,5,6,7]
[true]/[0]/[1]

Ã¢Â€ÂƒÃ¢Â†Â“
[7,6,5,4,3,2,1]

[true]/[0]/[1]

Ã¢Â€ÂƒÃ¢Â†Â“

[,]
[false,false]//

Ã¢Â€ÂƒÃ¢Â†Â“
[,]

edited Sep 3 at 13:54

asked Sep 3 at 12:44

AdÃ¡m

27.7k268185

I feel like the hardest part in most statically typed languages will be golfing the type signatures involved.
â€“Â ÃŽÂŸurous
Sep 3 at 22:07

@ÃŽÂŸurous How do those languages normally deal with arbitrary array data?
â€“Â AdÃ¡m
Sep 3 at 22:29

1

there's three cases for "normal" use as I see it: only worrying about one level of the array (eg: reverse works on arbitrary arrays but only cares about the first level), generics, or recursive classes (type / object classes depending on functional or OOP, but similar use-case). The latter two are usually far more verbose.
â€“Â ÃŽÂŸurous
Sep 3 at 22:39

Can we store the matrix as arrays of pointers to pointers (in C or asm), instead of proper multi-dimensional arrays where everything is contiguous in memory? I'm pretty sure all the normal higher-level / dynamically-typed languages with arbitrary nesting of lists are already treating things as lists of lists, not matrices, so I'm going to assume it's fine.
â€“Â Peter Cordes
Sep 5 at 3:42

@PeterCordes Sure, go ahead.
â€“Â AdÃ¡m
Sep 5 at 5:35

add a commentÂ |Â

up vote
23
down vote

favorite

Please state your input formats. Code explanations are much appreciated.

Walked-through example

We are given the 2-layer 3-row 4-column 3D-array

[[[ 1, 2, 3, 4],
 [ 5, 6, 7, 8],
 [ 9,10,11,12]],

 [[13,14,15,16],
 [17,18,19,20],
 [21,22,23,24]]]

and one of

[true,false,true] (Boolean list)
[0,2] (0-indexed list)
[1,3] (1-indexed list)

[[[13,14,15,16],
 [17,18,19,20],
 [21,22,23,24]],

 [[ 1, 2, 3, 4],
 [ 5, 6, 7, 8],
 [ 9,10,11,12]]]

and then we reverse the order of the elements of each row:

[[[16,15,14,13],
 [20,19,18,17],
 [24,23,22,21]],

 [[ 4, 3, 2, 1],
 [ 8, 7, 6, 5],
 [12,11,10, 9]]]

Test cases

[[1,2,3],[4,5,6]]
[true,false]/[0]/[1]

Ã¢Â€ÂƒÃ¢Â†Â“
[[4,5,6],[1,2,3]]

[[1],[4]]
[true,false]/[0]/[1]

Ã¢Â€ÂƒÃ¢Â†Â“
[[4],[1]]

[[7]]
[true,true]/[0,1]/[1,2]

Ã¢Â€ÂƒÃ¢Â†Â“
[[7]]

[1,2,3,4,5,6,7]
[true]/[0]/[1]

Ã¢Â€ÂƒÃ¢Â†Â“
[7,6,5,4,3,2,1]

[true]/[0]/[1]

Ã¢Â€ÂƒÃ¢Â†Â“

[,]
[false,false]//

Ã¢Â€ÂƒÃ¢Â†Â“
[,]

edited Sep 3 at 13:54

asked Sep 3 at 12:44

AdÃ¡m

27.7k268185

Please state your input formats. Code explanations are much appreciated.

Walked-through example

We are given the 2-layer 3-row 4-column 3D-array

[[[ 1, 2, 3, 4],
 [ 5, 6, 7, 8],
 [ 9,10,11,12]],

 [[13,14,15,16],
 [17,18,19,20],
 [21,22,23,24]]]

and one of

[true,false,true] (Boolean list)
[0,2] (0-indexed list)
[1,3] (1-indexed list)

[[[13,14,15,16],
 [17,18,19,20],
 [21,22,23,24]],

 [[ 1, 2, 3, 4],
 [ 5, 6, 7, 8],
 [ 9,10,11,12]]]

and then we reverse the order of the elements of each row:

[[[16,15,14,13],
 [20,19,18,17],
 [24,23,22,21]],

 [[ 4, 3, 2, 1],
 [ 8, 7, 6, 5],
 [12,11,10, 9]]]

Test cases

[[1,2,3],[4,5,6]]
[true,false]/[0]/[1]

Ã¢Â€ÂƒÃ¢Â†Â“
[[4,5,6],[1,2,3]]

[[1],[4]]
[true,false]/[0]/[1]

Ã¢Â€ÂƒÃ¢Â†Â“
[[4],[1]]

[[7]]
[true,true]/[0,1]/[1,2]

Ã¢Â€ÂƒÃ¢Â†Â“
[[7]]

[1,2,3,4,5,6,7]
[true]/[0]/[1]

Ã¢Â€ÂƒÃ¢Â†Â“
[7,6,5,4,3,2,1]

[true]/[0]/[1]

Ã¢Â€ÂƒÃ¢Â†Â“

[,]
[false,false]//

Ã¢Â€ÂƒÃ¢Â†Â“
[,]

edited Sep 3 at 13:54

asked Sep 3 at 12:44

AdÃ¡m

27.7k268185

edited Sep 3 at 13:54

asked Sep 3 at 12:44

AdÃ¡m

27.7k268185

asked Sep 3 at 12:44

AdÃ¡m

27.7k268185

asked Sep 3 at 12:44

AdÃ¡m

27.7k268185

I feel like the hardest part in most statically typed languages will be golfing the type signatures involved.
â€“Â ÃŽÂŸurous
Sep 3 at 22:07

@ÃŽÂŸurous How do those languages normally deal with arbitrary array data?
â€“Â AdÃ¡m
Sep 3 at 22:29

1

there's three cases for "normal" use as I see it: only worrying about one level of the array (eg: reverse works on arbitrary arrays but only cares about the first level), generics, or recursive classes (type / object classes depending on functional or OOP, but similar use-case). The latter two are usually far more verbose.
â€“Â ÃŽÂŸurous
Sep 3 at 22:39

Can we store the matrix as arrays of pointers to pointers (in C or asm), instead of proper multi-dimensional arrays where everything is contiguous in memory? I'm pretty sure all the normal higher-level / dynamically-typed languages with arbitrary nesting of lists are already treating things as lists of lists, not matrices, so I'm going to assume it's fine.
â€“Â Peter Cordes
Sep 5 at 3:42

@PeterCordes Sure, go ahead.
â€“Â AdÃ¡m
Sep 5 at 5:35

add a commentÂ |Â

I feel like the hardest part in most statically typed languages will be golfing the type signatures involved.
â€“Â ÃŽÂŸurous
Sep 3 at 22:07

@ÃŽÂŸurous How do those languages normally deal with arbitrary array data?
â€“Â AdÃ¡m
Sep 3 at 22:29

1

there's three cases for "normal" use as I see it: only worrying about one level of the array (eg: reverse works on arbitrary arrays but only cares about the first level), generics, or recursive classes (type / object classes depending on functional or OOP, but similar use-case). The latter two are usually far more verbose.
â€“Â ÃŽÂŸurous
Sep 3 at 22:39

Can we store the matrix as arrays of pointers to pointers (in C or asm), instead of proper multi-dimensional arrays where everything is contiguous in memory? I'm pretty sure all the normal higher-level / dynamically-typed languages with arbitrary nesting of lists are already treating things as lists of lists, not matrices, so I'm going to assume it's fine.
â€“Â Peter Cordes
Sep 5 at 3:42

@PeterCordes Sure, go ahead.
â€“Â AdÃ¡m
Sep 5 at 5:35

I feel like the hardest part in most statically typed languages will be golfing the type signatures involved.
â€“Â ÃŽÂŸurous
Sep 3 at 22:07

@ÃŽÂŸurous How do those languages normally deal with arbitrary array data?
â€“Â AdÃ¡m
Sep 3 at 22:29

there's three cases for "normal" use as I see it: only worrying about one level of the array (eg: reverse works on arbitrary arrays but only cares about the first level), generics, or recursive classes (type / object classes depending on functional or OOP, but similar use-case). The latter two are usually far more verbose.
â€“Â ÃŽÂŸurous
Sep 3 at 22:39

Can we store the matrix as arrays of pointers to pointers (in C or asm), instead of proper multi-dimensional arrays where everything is contiguous in memory? I'm pretty sure all the normal higher-level / dynamically-typed languages with arbitrary nesting of lists are already treating things as lists of lists, not matrices, so I'm going to assume it's fine.
â€“Â Peter Cordes
Sep 5 at 3:42

@PeterCordes Sure, go ahead.
â€“Â AdÃ¡m
Sep 5 at 5:35

add a commentÂ |Â

15 Answers
15

active

oldest

votes

up vote
8
down vote

APL (Dyalog), 20 9 bytes

Ã¢ÂŠÂƒÃ¢ÂŒÂ½[Ã¢ÂÂº]Ã¢ÂÂµ/

Try it online!

How?

/ - reduce - take the rightmost element in the input (the array) and apply the function with next left element as left argument

Ã¢ÂŒÂ½[Ã¢ÂÂº]Ã¢ÂÂµ - reverse in the left argument (Ã¢ÂÂº) dimension

Ã¢ÂŠÂƒ - flatten the enclosed array

edited Sep 3 at 18:07

answered Sep 3 at 13:05

Uriel

11.5k4936

add a commentÂ |Â

up vote
8
down vote

APL (Dyalog Unicode), 9 bytes

Ã¢ÂŠÂƒÃ¢ÂŠÂ–[Ã¢ÂÂº]Ã¢ÂÂµÃ¢ÂŒÂ¿

Try it online!

It looks like Uriel edited into something almost identical first, but I developed it independently. I thought this input format is invalid.

edited Sep 3 at 18:31

answered Sep 3 at 16:55

Erik the Outgolfer

29.4k42698

add a commentÂ |Â

up vote
7
down vote

JavaScript (Node.js), 58 55 53 45 bytes

Saved 8 bytes thanks to @Shaggy

Takes input as (indications)(array), where indications is a Boolean list.

f=([r,...b])=>a=>1/r?a.sort(_=>r).map(f(b)):a

Try it online!

Commented

f = ( // f is a function taking:
 [r, // r = next 'reverse' Boolean flag
 ...b] // b = array of remaining flags
) => // and returning an anonymous function taking:
 a => // a = array (or sub-array) to process, or atomic element
 1 / r ? // if r is defined:
 a.sort(_ => r) // reverse a if r = 1; leave it unchanged otherwise
 .map(f(b)) // for each element in the resulting array: do a recursive call,
 // using f to generate a new callback function for the next flag
 : // else:
 a // a must be an atomic element and is simply left unchanged

edited Sep 4 at 3:21

answered Sep 3 at 13:29

Arnauld

63.6k580268

Just using r inplace of r||-1 seems to work.
â€“Â Shaggy
Sep 3 at 15:18

Would f=([r,...b])=>a=>1/r?a.sort(_=>r).map(f(b)):a work? On my phone so can't test properly.
â€“Â Shaggy
Sep 3 at 22:10

@Shaggy Nice! I like this rather unusual processing flow.
â€“Â Arnauld
Sep 4 at 3:24

add a commentÂ |Â

up vote
6
down vote

Python 2, 56 55 bytes

f=lambda a,t:t and[f(l,t[1:])for l in a][::1|-t[0]]or a

Try it online!

edited Sep 3 at 13:57

answered Sep 3 at 13:02

TFeld

11.2k2833

add a commentÂ |Â

up vote
5
down vote

Jelly, 8 bytes

Ã¢Â€ÂÃ¢Â‚Â¬Ã¡ÂºÂ‹Ã¢Â€ÂÃ¡Â¹ÂšpFv

Takes a 0-indexed list of dimensions.

Try it online!

How it works

Ã¢Â€ÂÃ¢Â‚Â¬Ã¡ÂºÂ‹Ã¢Â€ÂÃ¡Â¹ÂšpFv Main link. Left arg: D (dimensions, 0-based), Right arg: A (array)

Ã¢Â€ÂÃ¢Â‚Â¬Ã¡ÂºÂ‹ Repeat 'Ã¢Â‚Â¬' d times, for each d in D.
 Ã¢Â€ÂÃ¡Â¹Âšp Perform Cartesian product of ['Ã¡Â¹Âš'] and each string of 'Ã¢Â‚Â¬'s, prepending a
 'Ã¡Â¹Âš' to each string of 'Ã¢Â‚Â¬'s.
 F Flatten the result.
 If, e.g., D = [0,2,4], we build the string "Ã¡Â¹ÂšÃ¡Â¹ÂšÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¡Â¹ÂšÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬".
 v Eval the resulting string, using A as left argument.

edited Sep 3 at 14:57

answered Sep 3 at 14:48

Dennisâ™¦

182k32291722

1

That's horrible. Very nice!
â€“Â AdÃ¡m
Sep 3 at 17:58

add a commentÂ |Â

up vote
5
down vote

R, 80 78 77 bytes

Create the call to R's extractor [ by creating a list of sequences reversed where indicated. They actually contain zeros, which are silently ignored. The drop=F is needed to prevent R's default dropping of dimensions. We need the rev call to the dimension reverse indicator, because of the way R fills arrays.

-2 thanks @Giuseppe

-1 using in-line assignment.

function(x,a,d=dim(x))do.call("[",c(list(x),Map(seq,r<-d*rev(a),d-r),drop=F))

Try it online!

Honorable mention to @JayCe who came up with a variation that gets the same result in the same length:

function(x,a,d=dim(x))array(x[t(t(expand.grid(Map(seq,r<-d*rev(a),d-r))))],d)

Try it online!

edited Sep 4 at 14:45

answered Sep 3 at 17:41

J.Doe

6018

1

78 bytes very nice answer!
â€“Â Giuseppe
Sep 3 at 19:38

1

Very profound answer. It took me a while to understand it completely. I tried replicating it without using do.call - it's longer at 83 bytes, still posting this here as a comment for reference: TIO
â€“Â JayCe
Sep 4 at 14:01

Well actually, @JayCe, your great answer can also be golfed to 78 bytes!
â€“Â J.Doe
Sep 4 at 14:11

add a commentÂ |Â

up vote
5
down vote

Haskell, 120 119 bytes

the function f takes the N-dimensional list and a list of bool as input

class F r where f::[Bool]->r->r
instance F Int where f=seq
instance F r=>F[r]where f(a:y)=last(id:[reverse|a]).map(f y)

edited Sep 6 at 6:18

answered Sep 3 at 14:43

Damien

2,379317

1

You don't need parentheses around F r.
â€“Â Ã˜rjan Johansen
Sep 3 at 18:01

1

TIO link with test cases and OJ's 1-byte golf.
â€“Â Scrooble
Sep 3 at 19:37

add a commentÂ |Â

up vote
4
down vote

05AB1E, 23 11 10 bytes

'Ã¢Â‚Â¬sÃƒÂ—'RÃ‚Â«J.V

Try it online.

-12 bytes thanks to @Mr.Xcoder.

Input as 0-indexed truthy-values (i.e. [0,2,3]), which is the first input.

Explanation:

'Ã¢Â‚Â¬sÃƒÂ— # Repeat "Ã¢Â‚Â¬" the indices amount of times
 # i.e. [0,2,3] Ã¢Â†Â’ ["","Ã¢Â‚Â¬Ã¢Â‚Â¬","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬"]
 'RÃ‚Â« # Append each with "R"
 # i.e. ["","Ã¢Â‚Â¬Ã¢Â‚Â¬","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬"] Ã¢Â†Â’ ["R","Ã¢Â‚Â¬Ã¢Â‚Â¬R","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R"]
 J # Join them all together
 # i.e. ["R","Ã¢Â‚Â¬Ã¢Â‚Â¬R","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R"] Ã¢Â†Â’ RÃ¢Â‚Â¬Ã¢Â‚Â¬RÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R
 .V # Execute string as 05AB1E code

For example: if the indices input-list is [0,2,3], it will create the following string:

RÃ¢Â‚Â¬Ã¢Â‚Â¬RÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R

Which will:

 Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R # Reverse the items in the most inner (4th level) lists
 Ã¢Â‚Â¬Ã¢Â‚Â¬R # Reverse the most inner (3rd level) lists themselves
 # Do nothing with the inner (2nd level) lists 
R # Reverse the entire outer (1st level) list

Original 23 byte answer:

Ã„Â‡URvyÃ¢Â€Âž RÃƒÂ¨JÃ¢Â€Â¦ÃŽÂµÃƒÂ¿}}Ã¢Â€Âž RXÃƒÂ¨J.V

Input as boolean-list (i.e. [1,0,1,1]), which is the first input.

Try it online.

Explanation:

Ã„Â‡U # Pop and save the first boolean in variable `X`
 R # Reverse the remaining boolean-list
 v } # Loop `y` over each of them:
 Ã¢Â€Âž RÃƒÂ¨ # Take the string "R ", and index the current boolean (0 or 1) in it
 J # Join it together with the string of the previous iteration
 Ã¢Â€Â¦ÃŽÂµÃƒÂ¿} # Surround it with "ÃŽÂµ" and "}"
 Ã¢Â€Âž RXÃƒÂ¨ # Index variable `X` also in "R "
 J # Join it together with the rest
.V # Execute string as 05AB1E code

For example: If the boolean input-list is [1,0,1,1], it will create the following string:

ÃŽÂµÃŽÂµÃŽÂµR}R} }R

Which will:

 ÃŽÂµR} # Reverse the items in the most inner (4th level) lists
 ÃŽÂµ R} # Reverse the most inner (3rd level) lists themselves
ÃŽÂµ } # Do nothing with the inner (2nd level) lists 
 R # Reverse the entire outer (1st level) list

edited Sep 3 at 17:10

answered Sep 3 at 13:03

Kevin Cruijssen

29.5k550162

1

Nice answer, but... uhm... would this 11 byter work?
â€“Â Mr. Xcoder
Sep 3 at 15:31

@Mr.Xcoder Thanks! That's indeed a lot easier. And been able to golf 1 more byte by appending each, instead of prepending and then reversing.
â€“Â Kevin Cruijssen
Sep 3 at 17:10

@Mr.Xcoder Btw, why does 'x* work to repeat x n amount of times without using a swap, but it doesn't work with 'Ã¢Â‚Â¬*?.. EDIT: Only in the legacy though..
â€“Â Kevin Cruijssen
Sep 3 at 17:12

The legacy version is quite buggy, it might be due to the fact that Ã¢Â‚Â¬ is still parsed as an operator even though it's in a char literal? Not sure to be honest. In the new version, * doesn't behave the same way nonetheless.
â€“Â Mr. Xcoder
Sep 3 at 17:18

add a commentÂ |Â

up vote
3
down vote

JavaScript (Node.js), 60 bytes

A different (recursive) approach. doesn't beat Arnauld's answer ... yet....

Takes the input as array, boolean list

f=(a,r)=>r>?(r[0]?a.reverse():a).map(c=>f(c,r.slice(1))):a

f=(a,r)=>r>?(r[0]?a.reverse():a).map(c=>f(c,r.slice(1))):a

console.log(f([[[[3,1,4,1],[5,9,2,6]],[[5,3,5,8],[9,7,9,3]]],[[[2,3,8,4],[6,2,6,4]],[[3,3,8,3],[2,7,9,5]]]],[true,false,true,true]))

edited Sep 3 at 14:03

answered Sep 3 at 13:21

Luis felipe De jesus Munoz

2,9761044

add a commentÂ |Â

up vote
3
down vote

Pyth, 15 bytes

.v+jk.n*_*LME

Try it here!

Annoyingly, handling the empty dimension list case takes no less than 2 bytes... I'd rather use ss in place of jk.n but :| Assumes that the list to be transformed can be given in native Pyth syntax, as a string. I've written a converter to Pyth syntax to make testing easier. In the unfortunate case that the OP chooses not to allow this, a 17-byter will "fix" it:

.v+jk.n*_*LMEE

edited Sep 3 at 16:10

answered Sep 3 at 16:05

Mr. Xcoder

30.3k758193

add a commentÂ |Â

up vote
3
down vote

Japt, 15 14 bytes

With some inspiration from Arnauld's solution.

Takes the indications as the first input, as a boolean array of 1s and 0s.

ÃƒÂŠ?Vn@ÃƒÂŽÃƒÂƒÃ‚Â£ÃƒÂŸUÃƒÂ…X:V

Try it

Explanation

 :Implicit input of boolean array U=indications and multi-dimensional integer array V
ÃƒÂŠ :Get the length of U
 ? :If truthy (i.e., >0)
 Vn : Sort V
 @ÃƒÂŽÃƒÂƒ : Function that gets the first element of U; 0 will leave the array untouched, 1 will reverse it.
 Ã‚Â£ : Map each X
 ÃƒÂŸ : Run the programme again with the following inputs
 UÃƒÂ… : U with the first element removed
 X : X will serve as the new value of V
 : :Else
 V : Just return V

edited Sep 3 at 16:12

answered Sep 3 at 15:38

Shaggy

16.3k21560

add a commentÂ |Â

up vote
3
down vote

Clean, 122 112 bytes

import StdEnv
class$r::[Bool]->r->r
instance$r where$_=id
instance$[r]| $r where$[a:y]=if(a)reverse id o map($y)

Try it online!

A version of Damien's Haskell answer using Clean's golfier type system. Really shows the extensive similarities between the two languages.

Explained:

import StdEnv // import basic stuff
class $ r :: [Bool] -> r -> r // $ takes a boolean list and returns a function on r to r
instance $ r // instance on all types, taken when no more specific instances exist
 where $ _ = id // return the identity function for all arguments
instance $ [r] | $ r // instance for lists of a type which has an instance itself
 where $ [a: y]
 = if(a) reverse id // reverse if the head of the argument is true
 o map ($ y) // composed with the map function acting on $ applied to the tail

edited Sep 4 at 1:37

answered Sep 3 at 22:33

ÃŽÂŸurous

5,2131931

add a commentÂ |Â

up vote
2
down vote

Ruby, 54 bytes

f=->a,dr,*z=d;r&&a.reverse!;d==?a:a.map

Try it online!

edited Sep 6 at 7:30

answered Sep 6 at 7:11

G B

6,3571324

add a commentÂ |Â

up vote
1
down vote

(untested but I think correct. compiler asm output looks like what I expect. Will update if/when I find time to write a test harness that creates and prints this data structure.)

GNU C++ (portable) 148 bytes

#include<algorithm>
#include<cstdint>
struct mintptr_t d,l,a;void R(int*r)if(*r)std::reverse(a,a+l);for(int i=0;d&&i<l;((m*)a[i++])->R(r+1));;

GNU C++ (int=pointer and falls off a non-void function UB) 120 bytes

#include<algorithm>
struct mint d,l,a,R(int*r)if(*r)std::reverse(a,a+l);for(int i=0;d&&i<l;((m*)a[i++])->R(r+1));;

This is a struct of depth counter, length, array of integers or pointers. In the bottom level of this non-binary tree (depth==0), the array of intptr_t is an array of integers. In higher levels, it's a struct m* stored in intptr_t. Traversal takes a cast.

The R() reverse function is a member function because that saves declaring an arg, and saves a lot of p-> syntax to reference the struct members vs. the implicit this pointer.

The only GNU extension is the C99 flexible array member to make a variable-sized struct, which is supported in C++ as a GNU extension. I could have used a *a member pointing to a separately-allocated array and have this be plain ISO C++. (And that would actually save a byte without requiring any other changes). I wrote this as a mockup / reference implementation for an asm version.

The shorter version with just int also declares R() as returning int instead of void. These two bits of hackery are unrelated; this is just the "works on at least one implementation" version.

It should work fine on 32-bit targets (where int can hold a pointer), as long as you compile with gcc7 or older, or disable optimizations. (gcc8 -O3 assumes that execution can't reach the bottom of a non-void function because that would be UB.) x86 gcc -m32 -O3 should work fine with gcc7, like on Godbolt where I included both versions (in different namespaces) and a non-member-function version.

Ungolfed

The function arg, int r, is an array of 0 / non-zero integers that indicate whether a given depth should be swapped, starting with the outer-most level.

#include<algorithm> // for std::reverse
#include<cstdint> // for intptr_t. GNU C defines __intptr_t, so we could use that...

struct m
 __intptr_t d,l,a; // depth = 0 means values, >0 means pointers.
 // l = length
 //__intptr_t a; // flexible array member: array contiguous with the struct

 void R(int r) 
 if(*r)
 std::reverse(a, a+l); // *r && std::reverse() doesn't work because it returns void.

 if(d) // recurse if this isn't the bottom depth
 for(int i=0 ; i<l ; i++) // tree traversal
 ((m*)a[i])->R(r+1); // with the rest of the depth list
 

; // struct m

When we recurse, we pass r+1, so checking the current depth is always *r.

An earlier version just passed r unchanged, and checked r[d]. With a flexible array member, I needed to store some kind of last-level indicator because a is not a pointer, it's a true array with no indirection. But with a intptr_t *a member, I couldn't just have that be nullptr for the leaf level, because I want it to be values.

Reversing the current level before or after the tree traversal shouldn't matter. I didn't try to do it during.

I'm not sure that std::reverse is worth the byte count vs. a manual loop, especially if I can work in calling R() on each pointer exactly once somewhere inside that loop. But only if d!=0

edited Sep 5 at 8:26

answered Sep 5 at 7:53

Peter Cordes

1,912917

Whoa, impressive.
â€“Â AdÃ¡m
Sep 5 at 8:08

@AdÃ¡m: thanks, it golfed down surprisingly naturally and well after I wrote it. I'm curious to see what I can do in x86 machine code :P
â€“Â Peter Cordes
Sep 5 at 8:29

add a commentÂ |Â

up vote
1
down vote

Mathematica, 7 bytes

Reverse

Function. Give it a nested list as the first argument, and the 1-based list of levels/dimensions to reverse as the second argument. Try it online!

^{^{Finally, another challenge where Mathematica has a builtin!}}

edited Sep 5 at 10:34

answered Sep 4 at 22:44

LegionMammal978

14.8k41752

layers to reverse? TIO link maybe?
â€“Â AdÃ¡m
Sep 5 at 1:52

@AdÃ¡m That is, the list of dimensions to reverse, generally referred to as levels in Mathematica.
â€“Â LegionMammal978
Sep 5 at 10:18

add a commentÂ |Â

Your Answer

StackExchange.ifUsing("editor", function ()
return StackExchange.using("mathjaxEditing", function ()
StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix)
StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["\$", "\$"]]);
);
);
, "mathjax-editing");

StackExchange.ifUsing("editor", function ()
StackExchange.using("externalEditor", function ()
StackExchange.using("snippets", function ()
StackExchange.snippets.init();
);
);
, "code-snippets");

StackExchange.ready(function()
var channelOptions =
tags: "".split(" "),
id: "200"
;
initTagRenderer("".split(" "), "".split(" "), channelOptions);

StackExchange.using("externalEditor", function()
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled)
StackExchange.using("snippets", function()
createEditor();
);

else
createEditor();

);

function createEditor()
StackExchange.prepareEditor(
heartbeatType: 'answer',
convertImagesToLinks: false,
noModals: false,
showLowRepImageUploadWarning: true,
reputationToPostImages: null,
bindNavPrevention: true,
postfix: "",
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
);

);

draft saved

draft discarded

StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fcodegolf.stackexchange.com%2fquestions%2f171627%2fmulti-dimensional-reversal%23new-answer', 'question_page');

);

Post as a guest

Name

15 Answers
15

active

oldest

votes

15 Answers
15

active

oldest

votes

up vote
8
down vote

APL (Dyalog), 20 9 bytes

Ã¢ÂŠÂƒÃ¢ÂŒÂ½[Ã¢ÂÂº]Ã¢ÂÂµ/

Try it online!

How?

/ - reduce - take the rightmost element in the input (the array) and apply the function with next left element as left argument

Ã¢ÂŒÂ½[Ã¢ÂÂº]Ã¢ÂÂµ - reverse in the left argument (Ã¢ÂÂº) dimension

Ã¢ÂŠÂƒ - flatten the enclosed array

edited Sep 3 at 18:07

answered Sep 3 at 13:05

Uriel

11.5k4936

add a commentÂ |Â

up vote
8
down vote

APL (Dyalog), 20 9 bytes

Ã¢ÂŠÂƒÃ¢ÂŒÂ½[Ã¢ÂÂº]Ã¢ÂÂµ/

Try it online!

How?

/ - reduce - take the rightmost element in the input (the array) and apply the function with next left element as left argument

Ã¢ÂŒÂ½[Ã¢ÂÂº]Ã¢ÂÂµ - reverse in the left argument (Ã¢ÂÂº) dimension

Ã¢ÂŠÂƒ - flatten the enclosed array

edited Sep 3 at 18:07

answered Sep 3 at 13:05

Uriel

11.5k4936

add a commentÂ |Â

up vote
8
down vote

APL (Dyalog), 20 9 bytes

Ã¢ÂŠÂƒÃ¢ÂŒÂ½[Ã¢ÂÂº]Ã¢ÂÂµ/

Try it online!

How?

/ - reduce - take the rightmost element in the input (the array) and apply the function with next left element as left argument

Ã¢ÂŒÂ½[Ã¢ÂÂº]Ã¢ÂÂµ - reverse in the left argument (Ã¢ÂÂº) dimension

Ã¢ÂŠÂƒ - flatten the enclosed array

edited Sep 3 at 18:07

answered Sep 3 at 13:05

Uriel

11.5k4936

APL (Dyalog), 20 9 bytes

Ã¢ÂŠÂƒÃ¢ÂŒÂ½[Ã¢ÂÂº]Ã¢ÂÂµ/

Try it online!

How?

/ - reduce - take the rightmost element in the input (the array) and apply the function with next left element as left argument

Ã¢ÂŒÂ½[Ã¢ÂÂº]Ã¢ÂÂµ - reverse in the left argument (Ã¢ÂÂº) dimension

Ã¢ÂŠÂƒ - flatten the enclosed array

edited Sep 3 at 18:07

answered Sep 3 at 13:05

Uriel

11.5k4936

edited Sep 3 at 18:07

answered Sep 3 at 13:05

Uriel

11.5k4936

answered Sep 3 at 13:05

Uriel

11.5k4936

answered Sep 3 at 13:05

Uriel

11.5k4936

add a commentÂ |Â

up vote
8
down vote

APL (Dyalog Unicode), 9 bytes

Ã¢ÂŠÂƒÃ¢ÂŠÂ–[Ã¢ÂÂº]Ã¢ÂÂµÃ¢ÂŒÂ¿

Try it online!

It looks like Uriel edited into something almost identical first, but I developed it independently. I thought this input format is invalid.

edited Sep 3 at 18:31

answered Sep 3 at 16:55

Erik the Outgolfer

29.4k42698

add a commentÂ |Â

up vote
8
down vote

APL (Dyalog Unicode), 9 bytes

Ã¢ÂŠÂƒÃ¢ÂŠÂ–[Ã¢ÂÂº]Ã¢ÂÂµÃ¢ÂŒÂ¿

Try it online!

It looks like Uriel edited into something almost identical first, but I developed it independently. I thought this input format is invalid.

edited Sep 3 at 18:31

answered Sep 3 at 16:55

Erik the Outgolfer

29.4k42698

add a commentÂ |Â

up vote
8
down vote

APL (Dyalog Unicode), 9 bytes

Ã¢ÂŠÂƒÃ¢ÂŠÂ–[Ã¢ÂÂº]Ã¢ÂÂµÃ¢ÂŒÂ¿

Try it online!

It looks like Uriel edited into something almost identical first, but I developed it independently. I thought this input format is invalid.

edited Sep 3 at 18:31

answered Sep 3 at 16:55

Erik the Outgolfer

29.4k42698

APL (Dyalog Unicode), 9 bytes

Ã¢ÂŠÂƒÃ¢ÂŠÂ–[Ã¢ÂÂº]Ã¢ÂÂµÃ¢ÂŒÂ¿

Try it online!

It looks like Uriel edited into something almost identical first, but I developed it independently. I thought this input format is invalid.

edited Sep 3 at 18:31

answered Sep 3 at 16:55

Erik the Outgolfer

29.4k42698

edited Sep 3 at 18:31

answered Sep 3 at 16:55

Erik the Outgolfer

29.4k42698

answered Sep 3 at 16:55

Erik the Outgolfer

29.4k42698

answered Sep 3 at 16:55

Erik the Outgolfer

29.4k42698

add a commentÂ |Â

up vote
7
down vote

JavaScript (Node.js), 58 55 53 45 bytes

Saved 8 bytes thanks to @Shaggy

Takes input as (indications)(array), where indications is a Boolean list.

f=([r,...b])=>a=>1/r?a.sort(_=>r).map(f(b)):a

Try it online!

Commented

f = ( // f is a function taking:
 [r, // r = next 'reverse' Boolean flag
 ...b] // b = array of remaining flags
) => // and returning an anonymous function taking:
 a => // a = array (or sub-array) to process, or atomic element
 1 / r ? // if r is defined:
 a.sort(_ => r) // reverse a if r = 1; leave it unchanged otherwise
 .map(f(b)) // for each element in the resulting array: do a recursive call,
 // using f to generate a new callback function for the next flag
 : // else:
 a // a must be an atomic element and is simply left unchanged

edited Sep 4 at 3:21

answered Sep 3 at 13:29

Arnauld

63.6k580268

Just using r inplace of r||-1 seems to work.
â€“Â Shaggy
Sep 3 at 15:18

Would f=([r,...b])=>a=>1/r?a.sort(_=>r).map(f(b)):a work? On my phone so can't test properly.
â€“Â Shaggy
Sep 3 at 22:10

@Shaggy Nice! I like this rather unusual processing flow.
â€“Â Arnauld
Sep 4 at 3:24

add a commentÂ |Â

up vote
7
down vote

JavaScript (Node.js), 58 55 53 45 bytes

Saved 8 bytes thanks to @Shaggy

Takes input as (indications)(array), where indications is a Boolean list.

f=([r,...b])=>a=>1/r?a.sort(_=>r).map(f(b)):a

Try it online!

Commented

f = ( // f is a function taking:
 [r, // r = next 'reverse' Boolean flag
 ...b] // b = array of remaining flags
) => // and returning an anonymous function taking:
 a => // a = array (or sub-array) to process, or atomic element
 1 / r ? // if r is defined:
 a.sort(_ => r) // reverse a if r = 1; leave it unchanged otherwise
 .map(f(b)) // for each element in the resulting array: do a recursive call,
 // using f to generate a new callback function for the next flag
 : // else:
 a // a must be an atomic element and is simply left unchanged

edited Sep 4 at 3:21

answered Sep 3 at 13:29

Arnauld

63.6k580268

Just using r inplace of r||-1 seems to work.
â€“Â Shaggy
Sep 3 at 15:18

Would f=([r,...b])=>a=>1/r?a.sort(_=>r).map(f(b)):a work? On my phone so can't test properly.
â€“Â Shaggy
Sep 3 at 22:10

@Shaggy Nice! I like this rather unusual processing flow.
â€“Â Arnauld
Sep 4 at 3:24

add a commentÂ |Â

up vote
7
down vote

JavaScript (Node.js), 58 55 53 45 bytes

Saved 8 bytes thanks to @Shaggy

Takes input as (indications)(array), where indications is a Boolean list.

f=([r,...b])=>a=>1/r?a.sort(_=>r).map(f(b)):a

Try it online!

Commented

f = ( // f is a function taking:
 [r, // r = next 'reverse' Boolean flag
 ...b] // b = array of remaining flags
) => // and returning an anonymous function taking:
 a => // a = array (or sub-array) to process, or atomic element
 1 / r ? // if r is defined:
 a.sort(_ => r) // reverse a if r = 1; leave it unchanged otherwise
 .map(f(b)) // for each element in the resulting array: do a recursive call,
 // using f to generate a new callback function for the next flag
 : // else:
 a // a must be an atomic element and is simply left unchanged

edited Sep 4 at 3:21

answered Sep 3 at 13:29

Arnauld

63.6k580268

JavaScript (Node.js), 58 55 53 45 bytes

Saved 8 bytes thanks to @Shaggy

Takes input as (indications)(array), where indications is a Boolean list.

f=([r,...b])=>a=>1/r?a.sort(_=>r).map(f(b)):a

Try it online!

Commented

f = ( // f is a function taking:
 [r, // r = next 'reverse' Boolean flag
 ...b] // b = array of remaining flags
) => // and returning an anonymous function taking:
 a => // a = array (or sub-array) to process, or atomic element
 1 / r ? // if r is defined:
 a.sort(_ => r) // reverse a if r = 1; leave it unchanged otherwise
 .map(f(b)) // for each element in the resulting array: do a recursive call,
 // using f to generate a new callback function for the next flag
 : // else:
 a // a must be an atomic element and is simply left unchanged

edited Sep 4 at 3:21

answered Sep 3 at 13:29

Arnauld

63.6k580268

edited Sep 4 at 3:21

answered Sep 3 at 13:29

Arnauld

63.6k580268

answered Sep 3 at 13:29

Arnauld

63.6k580268

answered Sep 3 at 13:29

Arnauld

63.6k580268

Just using r inplace of r||-1 seems to work.
â€“Â Shaggy
Sep 3 at 15:18

Would f=([r,...b])=>a=>1/r?a.sort(_=>r).map(f(b)):a work? On my phone so can't test properly.
â€“Â Shaggy
Sep 3 at 22:10

@Shaggy Nice! I like this rather unusual processing flow.
â€“Â Arnauld
Sep 4 at 3:24

add a commentÂ |Â

Just using r inplace of r||-1 seems to work.
â€“Â Shaggy
Sep 3 at 15:18

Would f=([r,...b])=>a=>1/r?a.sort(_=>r).map(f(b)):a work? On my phone so can't test properly.
â€“Â Shaggy
Sep 3 at 22:10

@Shaggy Nice! I like this rather unusual processing flow.
â€“Â Arnauld
Sep 4 at 3:24

Just using r inplace of r||-1 seems to work.
â€“Â Shaggy
Sep 3 at 15:18

Would f=([r,...b])=>a=>1/r?a.sort(_=>r).map(f(b)):a work? On my phone so can't test properly.
â€“Â Shaggy
Sep 3 at 22:10

@Shaggy Nice! I like this rather unusual processing flow.
â€“Â Arnauld
Sep 4 at 3:24

add a commentÂ |Â

up vote
6
down vote

Python 2, 56 55 bytes

f=lambda a,t:t and[f(l,t[1:])for l in a][::1|-t[0]]or a

Try it online!

edited Sep 3 at 13:57

answered Sep 3 at 13:02

TFeld

11.2k2833

add a commentÂ |Â

up vote
6
down vote

Python 2, 56 55 bytes

f=lambda a,t:t and[f(l,t[1:])for l in a][::1|-t[0]]or a

Try it online!

edited Sep 3 at 13:57

answered Sep 3 at 13:02

TFeld

11.2k2833

add a commentÂ |Â

up vote
6
down vote

Python 2, 56 55 bytes

f=lambda a,t:t and[f(l,t[1:])for l in a][::1|-t[0]]or a

Try it online!

edited Sep 3 at 13:57

answered Sep 3 at 13:02

TFeld

11.2k2833

Python 2, 56 55 bytes

f=lambda a,t:t and[f(l,t[1:])for l in a][::1|-t[0]]or a

Try it online!

edited Sep 3 at 13:57

answered Sep 3 at 13:02

TFeld

11.2k2833

edited Sep 3 at 13:57

answered Sep 3 at 13:02

TFeld

11.2k2833

answered Sep 3 at 13:02

TFeld

11.2k2833

answered Sep 3 at 13:02

TFeld

11.2k2833

add a commentÂ |Â

up vote
5
down vote

Jelly, 8 bytes

Ã¢Â€ÂÃ¢Â‚Â¬Ã¡ÂºÂ‹Ã¢Â€ÂÃ¡Â¹ÂšpFv

Takes a 0-indexed list of dimensions.

Try it online!

How it works

Ã¢Â€ÂÃ¢Â‚Â¬Ã¡ÂºÂ‹Ã¢Â€ÂÃ¡Â¹ÂšpFv Main link. Left arg: D (dimensions, 0-based), Right arg: A (array)

Ã¢Â€ÂÃ¢Â‚Â¬Ã¡ÂºÂ‹ Repeat 'Ã¢Â‚Â¬' d times, for each d in D.
 Ã¢Â€ÂÃ¡Â¹Âšp Perform Cartesian product of ['Ã¡Â¹Âš'] and each string of 'Ã¢Â‚Â¬'s, prepending a
 'Ã¡Â¹Âš' to each string of 'Ã¢Â‚Â¬'s.
 F Flatten the result.
 If, e.g., D = [0,2,4], we build the string "Ã¡Â¹ÂšÃ¡Â¹ÂšÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¡Â¹ÂšÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬".
 v Eval the resulting string, using A as left argument.

edited Sep 3 at 14:57

answered Sep 3 at 14:48

Dennisâ™¦

182k32291722

1

That's horrible. Very nice!
â€“Â AdÃ¡m
Sep 3 at 17:58

add a commentÂ |Â

up vote
5
down vote

Jelly, 8 bytes

Ã¢Â€ÂÃ¢Â‚Â¬Ã¡ÂºÂ‹Ã¢Â€ÂÃ¡Â¹ÂšpFv

Takes a 0-indexed list of dimensions.

Try it online!

How it works

Ã¢Â€ÂÃ¢Â‚Â¬Ã¡ÂºÂ‹Ã¢Â€ÂÃ¡Â¹ÂšpFv Main link. Left arg: D (dimensions, 0-based), Right arg: A (array)

Ã¢Â€ÂÃ¢Â‚Â¬Ã¡ÂºÂ‹ Repeat 'Ã¢Â‚Â¬' d times, for each d in D.
 Ã¢Â€ÂÃ¡Â¹Âšp Perform Cartesian product of ['Ã¡Â¹Âš'] and each string of 'Ã¢Â‚Â¬'s, prepending a
 'Ã¡Â¹Âš' to each string of 'Ã¢Â‚Â¬'s.
 F Flatten the result.
 If, e.g., D = [0,2,4], we build the string "Ã¡Â¹ÂšÃ¡Â¹ÂšÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¡Â¹ÂšÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬".
 v Eval the resulting string, using A as left argument.

edited Sep 3 at 14:57

answered Sep 3 at 14:48

Dennisâ™¦

182k32291722

1

That's horrible. Very nice!
â€“Â AdÃ¡m
Sep 3 at 17:58

add a commentÂ |Â

up vote
5
down vote

Jelly, 8 bytes

Ã¢Â€ÂÃ¢Â‚Â¬Ã¡ÂºÂ‹Ã¢Â€ÂÃ¡Â¹ÂšpFv

Takes a 0-indexed list of dimensions.

Try it online!

How it works

Ã¢Â€ÂÃ¢Â‚Â¬Ã¡ÂºÂ‹Ã¢Â€ÂÃ¡Â¹ÂšpFv Main link. Left arg: D (dimensions, 0-based), Right arg: A (array)

Ã¢Â€ÂÃ¢Â‚Â¬Ã¡ÂºÂ‹ Repeat 'Ã¢Â‚Â¬' d times, for each d in D.
 Ã¢Â€ÂÃ¡Â¹Âšp Perform Cartesian product of ['Ã¡Â¹Âš'] and each string of 'Ã¢Â‚Â¬'s, prepending a
 'Ã¡Â¹Âš' to each string of 'Ã¢Â‚Â¬'s.
 F Flatten the result.
 If, e.g., D = [0,2,4], we build the string "Ã¡Â¹ÂšÃ¡Â¹ÂšÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¡Â¹ÂšÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬".
 v Eval the resulting string, using A as left argument.

edited Sep 3 at 14:57

answered Sep 3 at 14:48

Dennisâ™¦

182k32291722

Jelly, 8 bytes

Ã¢Â€ÂÃ¢Â‚Â¬Ã¡ÂºÂ‹Ã¢Â€ÂÃ¡Â¹ÂšpFv

Takes a 0-indexed list of dimensions.

Try it online!

How it works

Ã¢Â€ÂÃ¢Â‚Â¬Ã¡ÂºÂ‹Ã¢Â€ÂÃ¡Â¹ÂšpFv Main link. Left arg: D (dimensions, 0-based), Right arg: A (array)

Ã¢Â€ÂÃ¢Â‚Â¬Ã¡ÂºÂ‹ Repeat 'Ã¢Â‚Â¬' d times, for each d in D.
 Ã¢Â€ÂÃ¡Â¹Âšp Perform Cartesian product of ['Ã¡Â¹Âš'] and each string of 'Ã¢Â‚Â¬'s, prepending a
 'Ã¡Â¹Âš' to each string of 'Ã¢Â‚Â¬'s.
 F Flatten the result.
 If, e.g., D = [0,2,4], we build the string "Ã¡Â¹ÂšÃ¡Â¹ÂšÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¡Â¹ÂšÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬".
 v Eval the resulting string, using A as left argument.

edited Sep 3 at 14:57

answered Sep 3 at 14:48

Dennisâ™¦

182k32291722

edited Sep 3 at 14:57

answered Sep 3 at 14:48

Dennisâ™¦

182k32291722

answered Sep 3 at 14:48

Dennisâ™¦

182k32291722

answered Sep 3 at 14:48

Dennisâ™¦

182k32291722

1

That's horrible. Very nice!
â€“Â AdÃ¡m
Sep 3 at 17:58

add a commentÂ |Â

1

That's horrible. Very nice!
â€“Â AdÃ¡m
Sep 3 at 17:58

That's horrible. Very nice!
â€“Â AdÃ¡m
Sep 3 at 17:58

add a commentÂ |Â

up vote
5
down vote

R, 80 78 77 bytes

-2 thanks @Giuseppe

-1 using in-line assignment.

function(x,a,d=dim(x))do.call("[",c(list(x),Map(seq,r<-d*rev(a),d-r),drop=F))

Try it online!

Honorable mention to @JayCe who came up with a variation that gets the same result in the same length:

function(x,a,d=dim(x))array(x[t(t(expand.grid(Map(seq,r<-d*rev(a),d-r))))],d)

Try it online!

edited Sep 4 at 14:45

answered Sep 3 at 17:41

J.Doe

6018

1

78 bytes very nice answer!
â€“Â Giuseppe
Sep 3 at 19:38

1

Very profound answer. It took me a while to understand it completely. I tried replicating it without using do.call - it's longer at 83 bytes, still posting this here as a comment for reference: TIO
â€“Â JayCe
Sep 4 at 14:01

Well actually, @JayCe, your great answer can also be golfed to 78 bytes!
â€“Â J.Doe
Sep 4 at 14:11

add a commentÂ |Â

up vote
5
down vote

R, 80 78 77 bytes

-2 thanks @Giuseppe

-1 using in-line assignment.

function(x,a,d=dim(x))do.call("[",c(list(x),Map(seq,r<-d*rev(a),d-r),drop=F))

Try it online!

Honorable mention to @JayCe who came up with a variation that gets the same result in the same length:

function(x,a,d=dim(x))array(x[t(t(expand.grid(Map(seq,r<-d*rev(a),d-r))))],d)

Try it online!

edited Sep 4 at 14:45

answered Sep 3 at 17:41

J.Doe

6018

1

78 bytes very nice answer!
â€“Â Giuseppe
Sep 3 at 19:38

1

Very profound answer. It took me a while to understand it completely. I tried replicating it without using do.call - it's longer at 83 bytes, still posting this here as a comment for reference: TIO
â€“Â JayCe
Sep 4 at 14:01

Well actually, @JayCe, your great answer can also be golfed to 78 bytes!
â€“Â J.Doe
Sep 4 at 14:11

add a commentÂ |Â

up vote
5
down vote

R, 80 78 77 bytes

-2 thanks @Giuseppe

-1 using in-line assignment.

function(x,a,d=dim(x))do.call("[",c(list(x),Map(seq,r<-d*rev(a),d-r),drop=F))

Try it online!

Honorable mention to @JayCe who came up with a variation that gets the same result in the same length:

function(x,a,d=dim(x))array(x[t(t(expand.grid(Map(seq,r<-d*rev(a),d-r))))],d)

Try it online!

edited Sep 4 at 14:45

answered Sep 3 at 17:41

J.Doe

6018

R, 80 78 77 bytes

-2 thanks @Giuseppe

-1 using in-line assignment.

function(x,a,d=dim(x))do.call("[",c(list(x),Map(seq,r<-d*rev(a),d-r),drop=F))

Try it online!

Honorable mention to @JayCe who came up with a variation that gets the same result in the same length:

function(x,a,d=dim(x))array(x[t(t(expand.grid(Map(seq,r<-d*rev(a),d-r))))],d)

Try it online!

edited Sep 4 at 14:45

answered Sep 3 at 17:41

J.Doe

6018

edited Sep 4 at 14:45

answered Sep 3 at 17:41

J.Doe

6018

answered Sep 3 at 17:41

J.Doe

6018

answered Sep 3 at 17:41

J.Doe

6018

1

78 bytes very nice answer!
â€“Â Giuseppe
Sep 3 at 19:38

1

Very profound answer. It took me a while to understand it completely. I tried replicating it without using do.call - it's longer at 83 bytes, still posting this here as a comment for reference: TIO
â€“Â JayCe
Sep 4 at 14:01

Well actually, @JayCe, your great answer can also be golfed to 78 bytes!
â€“Â J.Doe
Sep 4 at 14:11

add a commentÂ |Â

1

78 bytes very nice answer!
â€“Â Giuseppe
Sep 3 at 19:38

1

Very profound answer. It took me a while to understand it completely. I tried replicating it without using do.call - it's longer at 83 bytes, still posting this here as a comment for reference: TIO
â€“Â JayCe
Sep 4 at 14:01

Well actually, @JayCe, your great answer can also be golfed to 78 bytes!
â€“Â J.Doe
Sep 4 at 14:11

78 bytes very nice answer!
â€“Â Giuseppe
Sep 3 at 19:38

Very profound answer. It took me a while to understand it completely. I tried replicating it without using do.call - it's longer at 83 bytes, still posting this here as a comment for reference: TIO
â€“Â JayCe
Sep 4 at 14:01

Well actually, @JayCe, your great answer can also be golfed to 78 bytes!
â€“Â J.Doe
Sep 4 at 14:11

add a commentÂ |Â

up vote
5
down vote

Haskell, 120 119 bytes

the function f takes the N-dimensional list and a list of bool as input

class F r where f::[Bool]->r->r
instance F Int where f=seq
instance F r=>F[r]where f(a:y)=last(id:[reverse|a]).map(f y)

edited Sep 6 at 6:18

answered Sep 3 at 14:43

Damien

2,379317

1

You don't need parentheses around F r.
â€“Â Ã˜rjan Johansen
Sep 3 at 18:01

1

TIO link with test cases and OJ's 1-byte golf.
â€“Â Scrooble
Sep 3 at 19:37

add a commentÂ |Â

up vote
5
down vote

Haskell, 120 119 bytes

the function f takes the N-dimensional list and a list of bool as input

class F r where f::[Bool]->r->r
instance F Int where f=seq
instance F r=>F[r]where f(a:y)=last(id:[reverse|a]).map(f y)

edited Sep 6 at 6:18

answered Sep 3 at 14:43

Damien

2,379317

1

You don't need parentheses around F r.
â€“Â Ã˜rjan Johansen
Sep 3 at 18:01

1

TIO link with test cases and OJ's 1-byte golf.
â€“Â Scrooble
Sep 3 at 19:37

add a commentÂ |Â

up vote
5
down vote

Haskell, 120 119 bytes

the function f takes the N-dimensional list and a list of bool as input

class F r where f::[Bool]->r->r
instance F Int where f=seq
instance F r=>F[r]where f(a:y)=last(id:[reverse|a]).map(f y)

edited Sep 6 at 6:18

answered Sep 3 at 14:43

Damien

2,379317

Haskell, 120 119 bytes

the function f takes the N-dimensional list and a list of bool as input

class F r where f::[Bool]->r->r
instance F Int where f=seq
instance F r=>F[r]where f(a:y)=last(id:[reverse|a]).map(f y)

edited Sep 6 at 6:18

answered Sep 3 at 14:43

Damien

2,379317

edited Sep 6 at 6:18

answered Sep 3 at 14:43

Damien

2,379317

answered Sep 3 at 14:43

Damien

2,379317

answered Sep 3 at 14:43

Damien

2,379317

1

You don't need parentheses around F r.
â€“Â Ã˜rjan Johansen
Sep 3 at 18:01

1

TIO link with test cases and OJ's 1-byte golf.
â€“Â Scrooble
Sep 3 at 19:37

add a commentÂ |Â

1

You don't need parentheses around F r.
â€“Â Ã˜rjan Johansen
Sep 3 at 18:01

1

TIO link with test cases and OJ's 1-byte golf.
â€“Â Scrooble
Sep 3 at 19:37

You don't need parentheses around F r.
â€“Â Ã˜rjan Johansen
Sep 3 at 18:01

TIO link with test cases and OJ's 1-byte golf.
â€“Â Scrooble
Sep 3 at 19:37

add a commentÂ |Â

up vote
4
down vote

05AB1E, 23 11 10 bytes

'Ã¢Â‚Â¬sÃƒÂ—'RÃ‚Â«J.V

Try it online.

-12 bytes thanks to @Mr.Xcoder.

Input as 0-indexed truthy-values (i.e. [0,2,3]), which is the first input.

Explanation:

'Ã¢Â‚Â¬sÃƒÂ— # Repeat "Ã¢Â‚Â¬" the indices amount of times
 # i.e. [0,2,3] Ã¢Â†Â’ ["","Ã¢Â‚Â¬Ã¢Â‚Â¬","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬"]
 'RÃ‚Â« # Append each with "R"
 # i.e. ["","Ã¢Â‚Â¬Ã¢Â‚Â¬","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬"] Ã¢Â†Â’ ["R","Ã¢Â‚Â¬Ã¢Â‚Â¬R","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R"]
 J # Join them all together
 # i.e. ["R","Ã¢Â‚Â¬Ã¢Â‚Â¬R","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R"] Ã¢Â†Â’ RÃ¢Â‚Â¬Ã¢Â‚Â¬RÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R
 .V # Execute string as 05AB1E code

For example: if the indices input-list is [0,2,3], it will create the following string:

RÃ¢Â‚Â¬Ã¢Â‚Â¬RÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R

Which will:

 Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R # Reverse the items in the most inner (4th level) lists
 Ã¢Â‚Â¬Ã¢Â‚Â¬R # Reverse the most inner (3rd level) lists themselves
 # Do nothing with the inner (2nd level) lists 
R # Reverse the entire outer (1st level) list

Original 23 byte answer:

Ã„Â‡URvyÃ¢Â€Âž RÃƒÂ¨JÃ¢Â€Â¦ÃŽÂµÃƒÂ¿}}Ã¢Â€Âž RXÃƒÂ¨J.V

Input as boolean-list (i.e. [1,0,1,1]), which is the first input.

Try it online.

Explanation:

Ã„Â‡U # Pop and save the first boolean in variable `X`
 R # Reverse the remaining boolean-list
 v } # Loop `y` over each of them:
 Ã¢Â€Âž RÃƒÂ¨ # Take the string "R ", and index the current boolean (0 or 1) in it
 J # Join it together with the string of the previous iteration
 Ã¢Â€Â¦ÃŽÂµÃƒÂ¿} # Surround it with "ÃŽÂµ" and "}"
 Ã¢Â€Âž RXÃƒÂ¨ # Index variable `X` also in "R "
 J # Join it together with the rest
.V # Execute string as 05AB1E code

For example: If the boolean input-list is [1,0,1,1], it will create the following string:

ÃŽÂµÃŽÂµÃŽÂµR}R} }R

Which will:

 ÃŽÂµR} # Reverse the items in the most inner (4th level) lists
 ÃŽÂµ R} # Reverse the most inner (3rd level) lists themselves
ÃŽÂµ } # Do nothing with the inner (2nd level) lists 
 R # Reverse the entire outer (1st level) list

edited Sep 3 at 17:10

answered Sep 3 at 13:03

Kevin Cruijssen

29.5k550162

1

Nice answer, but... uhm... would this 11 byter work?
â€“Â Mr. Xcoder
Sep 3 at 15:31

@Mr.Xcoder Thanks! That's indeed a lot easier. And been able to golf 1 more byte by appending each, instead of prepending and then reversing.
â€“Â Kevin Cruijssen
Sep 3 at 17:10

@Mr.Xcoder Btw, why does 'x* work to repeat x n amount of times without using a swap, but it doesn't work with 'Ã¢Â‚Â¬*?.. EDIT: Only in the legacy though..
â€“Â Kevin Cruijssen
Sep 3 at 17:12

The legacy version is quite buggy, it might be due to the fact that Ã¢Â‚Â¬ is still parsed as an operator even though it's in a char literal? Not sure to be honest. In the new version, * doesn't behave the same way nonetheless.
â€“Â Mr. Xcoder
Sep 3 at 17:18

add a commentÂ |Â

up vote
4
down vote

05AB1E, 23 11 10 bytes

'Ã¢Â‚Â¬sÃƒÂ—'RÃ‚Â«J.V

Try it online.

-12 bytes thanks to @Mr.Xcoder.

Input as 0-indexed truthy-values (i.e. [0,2,3]), which is the first input.

Explanation:

'Ã¢Â‚Â¬sÃƒÂ— # Repeat "Ã¢Â‚Â¬" the indices amount of times
 # i.e. [0,2,3] Ã¢Â†Â’ ["","Ã¢Â‚Â¬Ã¢Â‚Â¬","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬"]
 'RÃ‚Â« # Append each with "R"
 # i.e. ["","Ã¢Â‚Â¬Ã¢Â‚Â¬","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬"] Ã¢Â†Â’ ["R","Ã¢Â‚Â¬Ã¢Â‚Â¬R","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R"]
 J # Join them all together
 # i.e. ["R","Ã¢Â‚Â¬Ã¢Â‚Â¬R","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R"] Ã¢Â†Â’ RÃ¢Â‚Â¬Ã¢Â‚Â¬RÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R
 .V # Execute string as 05AB1E code

For example: if the indices input-list is [0,2,3], it will create the following string:

RÃ¢Â‚Â¬Ã¢Â‚Â¬RÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R

Which will:

 Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R # Reverse the items in the most inner (4th level) lists
 Ã¢Â‚Â¬Ã¢Â‚Â¬R # Reverse the most inner (3rd level) lists themselves
 # Do nothing with the inner (2nd level) lists 
R # Reverse the entire outer (1st level) list

Original 23 byte answer:

Ã„Â‡URvyÃ¢Â€Âž RÃƒÂ¨JÃ¢Â€Â¦ÃŽÂµÃƒÂ¿}}Ã¢Â€Âž RXÃƒÂ¨J.V

Input as boolean-list (i.e. [1,0,1,1]), which is the first input.

Try it online.

Explanation:

Ã„Â‡U # Pop and save the first boolean in variable `X`
 R # Reverse the remaining boolean-list
 v } # Loop `y` over each of them:
 Ã¢Â€Âž RÃƒÂ¨ # Take the string "R ", and index the current boolean (0 or 1) in it
 J # Join it together with the string of the previous iteration
 Ã¢Â€Â¦ÃŽÂµÃƒÂ¿} # Surround it with "ÃŽÂµ" and "}"
 Ã¢Â€Âž RXÃƒÂ¨ # Index variable `X` also in "R "
 J # Join it together with the rest
.V # Execute string as 05AB1E code

For example: If the boolean input-list is [1,0,1,1], it will create the following string:

ÃŽÂµÃŽÂµÃŽÂµR}R} }R

Which will:

 ÃŽÂµR} # Reverse the items in the most inner (4th level) lists
 ÃŽÂµ R} # Reverse the most inner (3rd level) lists themselves
ÃŽÂµ } # Do nothing with the inner (2nd level) lists 
 R # Reverse the entire outer (1st level) list

edited Sep 3 at 17:10

answered Sep 3 at 13:03

Kevin Cruijssen

29.5k550162

1

Nice answer, but... uhm... would this 11 byter work?
â€“Â Mr. Xcoder
Sep 3 at 15:31

@Mr.Xcoder Thanks! That's indeed a lot easier. And been able to golf 1 more byte by appending each, instead of prepending and then reversing.
â€“Â Kevin Cruijssen
Sep 3 at 17:10

@Mr.Xcoder Btw, why does 'x* work to repeat x n amount of times without using a swap, but it doesn't work with 'Ã¢Â‚Â¬*?.. EDIT: Only in the legacy though..
â€“Â Kevin Cruijssen
Sep 3 at 17:12

The legacy version is quite buggy, it might be due to the fact that Ã¢Â‚Â¬ is still parsed as an operator even though it's in a char literal? Not sure to be honest. In the new version, * doesn't behave the same way nonetheless.
â€“Â Mr. Xcoder
Sep 3 at 17:18

add a commentÂ |Â

up vote
4
down vote

05AB1E, 23 11 10 bytes

'Ã¢Â‚Â¬sÃƒÂ—'RÃ‚Â«J.V

Try it online.

-12 bytes thanks to @Mr.Xcoder.

Input as 0-indexed truthy-values (i.e. [0,2,3]), which is the first input.

Explanation:

'Ã¢Â‚Â¬sÃƒÂ— # Repeat "Ã¢Â‚Â¬" the indices amount of times
 # i.e. [0,2,3] Ã¢Â†Â’ ["","Ã¢Â‚Â¬Ã¢Â‚Â¬","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬"]
 'RÃ‚Â« # Append each with "R"
 # i.e. ["","Ã¢Â‚Â¬Ã¢Â‚Â¬","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬"] Ã¢Â†Â’ ["R","Ã¢Â‚Â¬Ã¢Â‚Â¬R","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R"]
 J # Join them all together
 # i.e. ["R","Ã¢Â‚Â¬Ã¢Â‚Â¬R","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R"] Ã¢Â†Â’ RÃ¢Â‚Â¬Ã¢Â‚Â¬RÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R
 .V # Execute string as 05AB1E code

For example: if the indices input-list is [0,2,3], it will create the following string:

RÃ¢Â‚Â¬Ã¢Â‚Â¬RÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R

Which will:

 Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R # Reverse the items in the most inner (4th level) lists
 Ã¢Â‚Â¬Ã¢Â‚Â¬R # Reverse the most inner (3rd level) lists themselves
 # Do nothing with the inner (2nd level) lists 
R # Reverse the entire outer (1st level) list

Original 23 byte answer:

Ã„Â‡URvyÃ¢Â€Âž RÃƒÂ¨JÃ¢Â€Â¦ÃŽÂµÃƒÂ¿}}Ã¢Â€Âž RXÃƒÂ¨J.V

Input as boolean-list (i.e. [1,0,1,1]), which is the first input.

Try it online.

Explanation:

Ã„Â‡U # Pop and save the first boolean in variable `X`
 R # Reverse the remaining boolean-list
 v } # Loop `y` over each of them:
 Ã¢Â€Âž RÃƒÂ¨ # Take the string "R ", and index the current boolean (0 or 1) in it
 J # Join it together with the string of the previous iteration
 Ã¢Â€Â¦ÃŽÂµÃƒÂ¿} # Surround it with "ÃŽÂµ" and "}"
 Ã¢Â€Âž RXÃƒÂ¨ # Index variable `X` also in "R "
 J # Join it together with the rest
.V # Execute string as 05AB1E code

For example: If the boolean input-list is [1,0,1,1], it will create the following string:

ÃŽÂµÃŽÂµÃŽÂµR}R} }R

Which will:

 ÃŽÂµR} # Reverse the items in the most inner (4th level) lists
 ÃŽÂµ R} # Reverse the most inner (3rd level) lists themselves
ÃŽÂµ } # Do nothing with the inner (2nd level) lists 
 R # Reverse the entire outer (1st level) list

edited Sep 3 at 17:10

answered Sep 3 at 13:03

Kevin Cruijssen

29.5k550162

05AB1E, 23 11 10 bytes

'Ã¢Â‚Â¬sÃƒÂ—'RÃ‚Â«J.V

Try it online.

-12 bytes thanks to @Mr.Xcoder.

Input as 0-indexed truthy-values (i.e. [0,2,3]), which is the first input.

Explanation:

'Ã¢Â‚Â¬sÃƒÂ— # Repeat "Ã¢Â‚Â¬" the indices amount of times
 # i.e. [0,2,3] Ã¢Â†Â’ ["","Ã¢Â‚Â¬Ã¢Â‚Â¬","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬"]
 'RÃ‚Â« # Append each with "R"
 # i.e. ["","Ã¢Â‚Â¬Ã¢Â‚Â¬","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬"] Ã¢Â†Â’ ["R","Ã¢Â‚Â¬Ã¢Â‚Â¬R","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R"]
 J # Join them all together
 # i.e. ["R","Ã¢Â‚Â¬Ã¢Â‚Â¬R","Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R"] Ã¢Â†Â’ RÃ¢Â‚Â¬Ã¢Â‚Â¬RÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R
 .V # Execute string as 05AB1E code

For example: if the indices input-list is [0,2,3], it will create the following string:

RÃ¢Â‚Â¬Ã¢Â‚Â¬RÃ¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R

Which will:

 Ã¢Â‚Â¬Ã¢Â‚Â¬Ã¢Â‚Â¬R # Reverse the items in the most inner (4th level) lists
 Ã¢Â‚Â¬Ã¢Â‚Â¬R # Reverse the most inner (3rd level) lists themselves
 # Do nothing with the inner (2nd level) lists 
R # Reverse the entire outer (1st level) list

Original 23 byte answer:

Ã„Â‡URvyÃ¢Â€Âž RÃƒÂ¨JÃ¢Â€Â¦ÃŽÂµÃƒÂ¿}}Ã¢Â€Âž RXÃƒÂ¨J.V

Input as boolean-list (i.e. [1,0,1,1]), which is the first input.

Try it online.

Explanation:

Ã„Â‡U # Pop and save the first boolean in variable `X`
 R # Reverse the remaining boolean-list
 v } # Loop `y` over each of them:
 Ã¢Â€Âž RÃƒÂ¨ # Take the string "R ", and index the current boolean (0 or 1) in it
 J # Join it together with the string of the previous iteration
 Ã¢Â€Â¦ÃŽÂµÃƒÂ¿} # Surround it with "ÃŽÂµ" and "}"
 Ã¢Â€Âž RXÃƒÂ¨ # Index variable `X` also in "R "
 J # Join it together with the rest
.V # Execute string as 05AB1E code

For example: If the boolean input-list is [1,0,1,1], it will create the following string:

ÃŽÂµÃŽÂµÃŽÂµR}R} }R

Which will:

 ÃŽÂµR} # Reverse the items in the most inner (4th level) lists
 ÃŽÂµ R} # Reverse the most inner (3rd level) lists themselves
ÃŽÂµ } # Do nothing with the inner (2nd level) lists 
 R # Reverse the entire outer (1st level) list

edited Sep 3 at 17:10

answered Sep 3 at 13:03

Kevin Cruijssen

29.5k550162

edited Sep 3 at 17:10

answered Sep 3 at 13:03

Kevin Cruijssen

29.5k550162

answered Sep 3 at 13:03

Kevin Cruijssen

29.5k550162

answered Sep 3 at 13:03

Kevin Cruijssen

29.5k550162

1

Nice answer, but... uhm... would this 11 byter work?
â€“Â Mr. Xcoder
Sep 3 at 15:31

@Mr.Xcoder Thanks! That's indeed a lot easier. And been able to golf 1 more byte by appending each, instead of prepending and then reversing.
â€“Â Kevin Cruijssen
Sep 3 at 17:10

@Mr.Xcoder Btw, why does 'x* work to repeat x n amount of times without using a swap, but it doesn't work with 'Ã¢Â‚Â¬*?.. EDIT: Only in the legacy though..
â€“Â Kevin Cruijssen
Sep 3 at 17:12

The legacy version is quite buggy, it might be due to the fact that Ã¢Â‚Â¬ is still parsed as an operator even though it's in a char literal? Not sure to be honest. In the new version, * doesn't behave the same way nonetheless.
â€“Â Mr. Xcoder
Sep 3 at 17:18

add a commentÂ |Â

1

Nice answer, but... uhm... would this 11 byter work?
â€“Â Mr. Xcoder
Sep 3 at 15:31

@Mr.Xcoder Thanks! That's indeed a lot easier. And been able to golf 1 more byte by appending each, instead of prepending and then reversing.
â€“Â Kevin Cruijssen
Sep 3 at 17:10

@Mr.Xcoder Btw, why does 'x* work to repeat x n amount of times without using a swap, but it doesn't work with 'Ã¢Â‚Â¬*?.. EDIT: Only in the legacy though..
â€“Â Kevin Cruijssen
Sep 3 at 17:12

The legacy version is quite buggy, it might be due to the fact that Ã¢Â‚Â¬ is still parsed as an operator even though it's in a char literal? Not sure to be honest. In the new version, * doesn't behave the same way nonetheless.
â€“Â Mr. Xcoder
Sep 3 at 17:18

Nice answer, but... uhm... would this 11 byter work?
â€“Â Mr. Xcoder
Sep 3 at 15:31

@Mr.Xcoder Thanks! That's indeed a lot easier. And been able to golf 1 more byte by appending each, instead of prepending and then reversing.
â€“Â Kevin Cruijssen
Sep 3 at 17:10

@Mr.Xcoder Btw, why does 'x* work to repeat x n amount of times without using a swap, but it doesn't work with 'Ã¢Â‚Â¬*?.. EDIT: Only in the legacy though..
â€“Â Kevin Cruijssen
Sep 3 at 17:12

The legacy version is quite buggy, it might be due to the fact that Ã¢Â‚Â¬ is still parsed as an operator even though it's in a char literal? Not sure to be honest. In the new version, * doesn't behave the same way nonetheless.
â€“Â Mr. Xcoder
Sep 3 at 17:18

add a commentÂ |Â

up vote
3
down vote

JavaScript (Node.js), 60 bytes

A different (recursive) approach. doesn't beat Arnauld's answer ... yet....

Takes the input as array, boolean list

f=(a,r)=>r>?(r[0]?a.reverse():a).map(c=>f(c,r.slice(1))):a

f=(a,r)=>r>?(r[0]?a.reverse():a).map(c=>f(c,r.slice(1))):a

console.log(f([[[[3,1,4,1],[5,9,2,6]],[[5,3,5,8],[9,7,9,3]]],[[[2,3,8,4],[6,2,6,4]],[[3,3,8,3],[2,7,9,5]]]],[true,false,true,true]))

edited Sep 3 at 14:03

answered Sep 3 at 13:21

Luis felipe De jesus Munoz

2,9761044

add a commentÂ |Â

up vote
3
down vote

JavaScript (Node.js), 60 bytes

A different (recursive) approach. doesn't beat Arnauld's answer ... yet....

Takes the input as array, boolean list

f=(a,r)=>r>?(r[0]?a.reverse():a).map(c=>f(c,r.slice(1))):a

f=(a,r)=>r>?(r[0]?a.reverse():a).map(c=>f(c,r.slice(1))):a

console.log(f([[[[3,1,4,1],[5,9,2,6]],[[5,3,5,8],[9,7,9,3]]],[[[2,3,8,4],[6,2,6,4]],[[3,3,8,3],[2,7,9,5]]]],[true,false,true,true]))

edited Sep 3 at 14:03

answered Sep 3 at 13:21

Luis felipe De jesus Munoz

2,9761044

add a commentÂ |Â

up vote
3
down vote

JavaScript (Node.js), 60 bytes

A different (recursive) approach. doesn't beat Arnauld's answer ... yet....

Takes the input as array, boolean list

f=(a,r)=>r>?(r[0]?a.reverse():a).map(c=>f(c,r.slice(1))):a

f=(a,r)=>r>?(r[0]?a.reverse():a).map(c=>f(c,r.slice(1))):a

console.log(f([[[[3,1,4,1],[5,9,2,6]],[[5,3,5,8],[9,7,9,3]]],[[[2,3,8,4],[6,2,6,4]],[[3,3,8,3],[2,7,9,5]]]],[true,false,true,true]))

edited Sep 3 at 14:03

answered Sep 3 at 13:21

Luis felipe De jesus Munoz

2,9761044

JavaScript (Node.js), 60 bytes

A different (recursive) approach. doesn't beat Arnauld's answer ... yet....

Takes the input as array, boolean list

f=(a,r)=>r>?(r[0]?a.reverse():a).map(c=>f(c,r.slice(1))):a

f=(a,r)=>r>?(r[0]?a.reverse():a).map(c=>f(c,r.slice(1))):a

console.log(f([[[[3,1,4,1],[5,9,2,6]],[[5,3,5,8],[9,7,9,3]]],[[[2,3,8,4],[6,2,6,4]],[[3,3,8,3],[2,7,9,5]]]],[true,false,true,true]))

f=(a,r)=>r>?(r[0]?a.reverse():a).map(c=>f(c,r.slice(1))):a

console.log(f([[[[3,1,4,1],[5,9,2,6]],[[5,3,5,8],[9,7,9,3]]],[[[2,3,8,4],[6,2,6,4]],[[3,3,8,3],[2,7,9,5]]]],[true,false,true,true]))

f=(a,r)=>r>?(r[0]?a.reverse():a).map(c=>f(c,r.slice(1))):a

console.log(f([[[[3,1,4,1],[5,9,2,6]],[[5,3,5,8],[9,7,9,3]]],[[[2,3,8,4],[6,2,6,4]],[[3,3,8,3],[2,7,9,5]]]],[true,false,true,true]))

edited Sep 3 at 14:03

answered Sep 3 at 13:21

Luis felipe De jesus Munoz

2,9761044

edited Sep 3 at 14:03

answered Sep 3 at 13:21

Luis felipe De jesus Munoz

2,9761044

answered Sep 3 at 13:21

Luis felipe De jesus Munoz

2,9761044

answered Sep 3 at 13:21

Luis felipe De jesus Munoz

2,9761044

add a commentÂ |Â

up vote
3
down vote

Pyth, 15 bytes

.v+jk.n*_*LME

Try it here!

.v+jk.n*_*LMEE

edited Sep 3 at 16:10

answered Sep 3 at 16:05

Mr. Xcoder

30.3k758193

add a commentÂ |Â

up vote
3
down vote

Pyth, 15 bytes

.v+jk.n*_*LME

Try it here!

.v+jk.n*_*LMEE

edited Sep 3 at 16:10

answered Sep 3 at 16:05

Mr. Xcoder

30.3k758193

add a commentÂ |Â

up vote
3
down vote

Pyth, 15 bytes

.v+jk.n*_*LME

Try it here!

.v+jk.n*_*LMEE

edited Sep 3 at 16:10

answered Sep 3 at 16:05

Mr. Xcoder

30.3k758193

Pyth, 15 bytes

.v+jk.n*_*LME

Try it here!

.v+jk.n*_*LMEE

edited Sep 3 at 16:10

answered Sep 3 at 16:05

Mr. Xcoder

30.3k758193

edited Sep 3 at 16:10

answered Sep 3 at 16:05

Mr. Xcoder

30.3k758193

answered Sep 3 at 16:05

Mr. Xcoder

30.3k758193

answered Sep 3 at 16:05

Mr. Xcoder

30.3k758193

add a commentÂ |Â

up vote
3
down vote

Japt, 15 14 bytes

With some inspiration from Arnauld's solution.

Takes the indications as the first input, as a boolean array of 1s and 0s.

ÃƒÂŠ?Vn@ÃƒÂŽÃƒÂƒÃ‚Â£ÃƒÂŸUÃƒÂ…X:V

Try it

Explanation

 :Implicit input of boolean array U=indications and multi-dimensional integer array V
ÃƒÂŠ :Get the length of U
 ? :If truthy (i.e., >0)
 Vn : Sort V
 @ÃƒÂŽÃƒÂƒ : Function that gets the first element of U; 0 will leave the array untouched, 1 will reverse it.
 Ã‚Â£ : Map each X
 ÃƒÂŸ : Run the programme again with the following inputs
 UÃƒÂ… : U with the first element removed
 X : X will serve as the new value of V
 : :Else
 V : Just return V

edited Sep 3 at 16:12

answered Sep 3 at 15:38

Shaggy

16.3k21560

add a commentÂ |Â

up vote
3
down vote

Japt, 15 14 bytes

With some inspiration from Arnauld's solution.

Takes the indications as the first input, as a boolean array of 1s and 0s.

ÃƒÂŠ?Vn@ÃƒÂŽÃƒÂƒÃ‚Â£ÃƒÂŸUÃƒÂ…X:V

Try it

Explanation

 :Implicit input of boolean array U=indications and multi-dimensional integer array V
ÃƒÂŠ :Get the length of U
 ? :If truthy (i.e., >0)
 Vn : Sort V
 @ÃƒÂŽÃƒÂƒ : Function that gets the first element of U; 0 will leave the array untouched, 1 will reverse it.
 Ã‚Â£ : Map each X
 ÃƒÂŸ : Run the programme again with the following inputs
 UÃƒÂ… : U with the first element removed
 X : X will serve as the new value of V
 : :Else
 V : Just return V

edited Sep 3 at 16:12

answered Sep 3 at 15:38

Shaggy

16.3k21560

add a commentÂ |Â

up vote
3
down vote

Japt, 15 14 bytes

With some inspiration from Arnauld's solution.

Takes the indications as the first input, as a boolean array of 1s and 0s.

ÃƒÂŠ?Vn@ÃƒÂŽÃƒÂƒÃ‚Â£ÃƒÂŸUÃƒÂ…X:V

Try it

Explanation

 :Implicit input of boolean array U=indications and multi-dimensional integer array V
ÃƒÂŠ :Get the length of U
 ? :If truthy (i.e., >0)
 Vn : Sort V
 @ÃƒÂŽÃƒÂƒ : Function that gets the first element of U; 0 will leave the array untouched, 1 will reverse it.
 Ã‚Â£ : Map each X
 ÃƒÂŸ : Run the programme again with the following inputs
 UÃƒÂ… : U with the first element removed
 X : X will serve as the new value of V
 : :Else
 V : Just return V

edited Sep 3 at 16:12

answered Sep 3 at 15:38

Shaggy

16.3k21560

Japt, 15 14 bytes

With some inspiration from Arnauld's solution.

Takes the indications as the first input, as a boolean array of 1s and 0s.

ÃƒÂŠ?Vn@ÃƒÂŽÃƒÂƒÃ‚Â£ÃƒÂŸUÃƒÂ…X:V

Try it

Explanation

 :Implicit input of boolean array U=indications and multi-dimensional integer array V
ÃƒÂŠ :Get the length of U
 ? :If truthy (i.e., >0)
 Vn : Sort V
 @ÃƒÂŽÃƒÂƒ : Function that gets the first element of U; 0 will leave the array untouched, 1 will reverse it.
 Ã‚Â£ : Map each X
 ÃƒÂŸ : Run the programme again with the following inputs
 UÃƒÂ… : U with the first element removed
 X : X will serve as the new value of V
 : :Else
 V : Just return V

edited Sep 3 at 16:12

answered Sep 3 at 15:38

Shaggy

16.3k21560

edited Sep 3 at 16:12

answered Sep 3 at 15:38

Shaggy

16.3k21560

answered Sep 3 at 15:38

Shaggy

16.3k21560

answered Sep 3 at 15:38

Shaggy

16.3k21560

add a commentÂ |Â

up vote
3
down vote

Clean, 122 112 bytes

import StdEnv
class$r::[Bool]->r->r
instance$r where$_=id
instance$[r]| $r where$[a:y]=if(a)reverse id o map($y)

Try it online!

A version of Damien's Haskell answer using Clean's golfier type system. Really shows the extensive similarities between the two languages.

Explained:

import StdEnv // import basic stuff
class $ r :: [Bool] -> r -> r // $ takes a boolean list and returns a function on r to r
instance $ r // instance on all types, taken when no more specific instances exist
 where $ _ = id // return the identity function for all arguments
instance $ [r] | $ r // instance for lists of a type which has an instance itself
 where $ [a: y]
 = if(a) reverse id // reverse if the head of the argument is true
 o map ($ y) // composed with the map function acting on $ applied to the tail

edited Sep 4 at 1:37

answered Sep 3 at 22:33

ÃŽÂŸurous

5,2131931

add a commentÂ |Â

up vote
3
down vote

Clean, 122 112 bytes

import StdEnv
class$r::[Bool]->r->r
instance$r where$_=id
instance$[r]| $r where$[a:y]=if(a)reverse id o map($y)

Try it online!

A version of Damien's Haskell answer using Clean's golfier type system. Really shows the extensive similarities between the two languages.

Explained:

import StdEnv // import basic stuff
class $ r :: [Bool] -> r -> r // $ takes a boolean list and returns a function on r to r
instance $ r // instance on all types, taken when no more specific instances exist
 where $ _ = id // return the identity function for all arguments
instance $ [r] | $ r // instance for lists of a type which has an instance itself
 where $ [a: y]
 = if(a) reverse id // reverse if the head of the argument is true
 o map ($ y) // composed with the map function acting on $ applied to the tail

edited Sep 4 at 1:37

answered Sep 3 at 22:33

ÃŽÂŸurous

5,2131931

add a commentÂ |Â

up vote
3
down vote

Clean, 122 112 bytes

import StdEnv
class$r::[Bool]->r->r
instance$r where$_=id
instance$[r]| $r where$[a:y]=if(a)reverse id o map($y)

Try it online!

A version of Damien's Haskell answer using Clean's golfier type system. Really shows the extensive similarities between the two languages.

Explained:

import StdEnv // import basic stuff
class $ r :: [Bool] -> r -> r // $ takes a boolean list and returns a function on r to r
instance $ r // instance on all types, taken when no more specific instances exist
 where $ _ = id // return the identity function for all arguments
instance $ [r] | $ r // instance for lists of a type which has an instance itself
 where $ [a: y]
 = if(a) reverse id // reverse if the head of the argument is true
 o map ($ y) // composed with the map function acting on $ applied to the tail

edited Sep 4 at 1:37

answered Sep 3 at 22:33

ÃŽÂŸurous

5,2131931

Clean, 122 112 bytes

import StdEnv
class$r::[Bool]->r->r
instance$r where$_=id
instance$[r]| $r where$[a:y]=if(a)reverse id o map($y)

Try it online!

A version of Damien's Haskell answer using Clean's golfier type system. Really shows the extensive similarities between the two languages.

Explained:

import StdEnv // import basic stuff
class $ r :: [Bool] -> r -> r // $ takes a boolean list and returns a function on r to r
instance $ r // instance on all types, taken when no more specific instances exist
 where $ _ = id // return the identity function for all arguments
instance $ [r] | $ r // instance for lists of a type which has an instance itself
 where $ [a: y]
 = if(a) reverse id // reverse if the head of the argument is true
 o map ($ y) // composed with the map function acting on $ applied to the tail

edited Sep 4 at 1:37

answered Sep 3 at 22:33

ÃŽÂŸurous

5,2131931

edited Sep 4 at 1:37

answered Sep 3 at 22:33

ÃŽÂŸurous

5,2131931

answered Sep 3 at 22:33

ÃŽÂŸurous

5,2131931

answered Sep 3 at 22:33

ÃŽÂŸurous

5,2131931

add a commentÂ |Â

up vote
2
down vote

Ruby, 54 bytes

f=->a,dr,*z=d;r&&a.reverse!;d==?a:a.map

Try it online!

edited Sep 6 at 7:30

answered Sep 6 at 7:11

G B

6,3571324

add a commentÂ |Â

up vote
2
down vote

Ruby, 54 bytes

f=->a,dr,*z=d;r&&a.reverse!;d==?a:a.map

Try it online!

edited Sep 6 at 7:30

answered Sep 6 at 7:11

G B

6,3571324

add a commentÂ |Â

up vote
2
down vote

Ruby, 54 bytes

f=->a,dr,*z=d;r&&a.reverse!;d==?a:a.map

Try it online!

edited Sep 6 at 7:30

answered Sep 6 at 7:11

G B

6,3571324

Ruby, 54 bytes

f=->a,dr,*z=d;r&&a.reverse!;d==?a:a.map

Try it online!

edited Sep 6 at 7:30

answered Sep 6 at 7:11

G B

6,3571324

edited Sep 6 at 7:30

answered Sep 6 at 7:11

G B

6,3571324

answered Sep 6 at 7:11

G B

6,3571324

answered Sep 6 at 7:11

G B

6,3571324

add a commentÂ |Â

up vote
1
down vote

(untested but I think correct. compiler asm output looks like what I expect. Will update if/when I find time to write a test harness that creates and prints this data structure.)

GNU C++ (portable) 148 bytes

#include<algorithm>
#include<cstdint>
struct mintptr_t d,l,a;void R(int*r)if(*r)std::reverse(a,a+l);for(int i=0;d&&i<l;((m*)a[i++])->R(r+1));;

GNU C++ (int=pointer and falls off a non-void function UB) 120 bytes

#include<algorithm>
struct mint d,l,a,R(int*r)if(*r)std::reverse(a,a+l);for(int i=0;d&&i<l;((m*)a[i++])->R(r+1));;

The R() reverse function is a member function because that saves declaring an arg, and saves a lot of p-> syntax to reference the struct members vs. the implicit this pointer.

The shorter version with just int also declares R() as returning int instead of void. These two bits of hackery are unrelated; this is just the "works on at least one implementation" version.

Ungolfed

The function arg, int r, is an array of 0 / non-zero integers that indicate whether a given depth should be swapped, starting with the outer-most level.

#include<algorithm> // for std::reverse
#include<cstdint> // for intptr_t. GNU C defines __intptr_t, so we could use that...

struct m
 __intptr_t d,l,a; // depth = 0 means values, >0 means pointers.
 // l = length
 //__intptr_t a; // flexible array member: array contiguous with the struct

 void R(int r) 
 if(*r)
 std::reverse(a, a+l); // *r && std::reverse() doesn't work because it returns void.

 if(d) // recurse if this isn't the bottom depth
 for(int i=0 ; i<l ; i++) // tree traversal
 ((m*)a[i])->R(r+1); // with the rest of the depth list
 

; // struct m

When we recurse, we pass r+1, so checking the current depth is always *r.

Reversing the current level before or after the tree traversal shouldn't matter. I didn't try to do it during.

I'm not sure that std::reverse is worth the byte count vs. a manual loop, especially if I can work in calling R() on each pointer exactly once somewhere inside that loop. But only if d!=0

edited Sep 5 at 8:26

answered Sep 5 at 7:53

Peter Cordes

1,912917

Whoa, impressive.
â€“Â AdÃ¡m
Sep 5 at 8:08

@AdÃ¡m: thanks, it golfed down surprisingly naturally and well after I wrote it. I'm curious to see what I can do in x86 machine code :P
â€“Â Peter Cordes
Sep 5 at 8:29

add a commentÂ |Â

up vote
1
down vote

(untested but I think correct. compiler asm output looks like what I expect. Will update if/when I find time to write a test harness that creates and prints this data structure.)

GNU C++ (portable) 148 bytes

#include<algorithm>
#include<cstdint>
struct mintptr_t d,l,a;void R(int*r)if(*r)std::reverse(a,a+l);for(int i=0;d&&i<l;((m*)a[i++])->R(r+1));;

GNU C++ (int=pointer and falls off a non-void function UB) 120 bytes

#include<algorithm>
struct mint d,l,a,R(int*r)if(*r)std::reverse(a,a+l);for(int i=0;d&&i<l;((m*)a[i++])->R(r+1));;

The R() reverse function is a member function because that saves declaring an arg, and saves a lot of p-> syntax to reference the struct members vs. the implicit this pointer.

The shorter version with just int also declares R() as returning int instead of void. These two bits of hackery are unrelated; this is just the "works on at least one implementation" version.

Ungolfed

The function arg, int r, is an array of 0 / non-zero integers that indicate whether a given depth should be swapped, starting with the outer-most level.

#include<algorithm> // for std::reverse
#include<cstdint> // for intptr_t. GNU C defines __intptr_t, so we could use that...

struct m
 __intptr_t d,l,a; // depth = 0 means values, >0 means pointers.
 // l = length
 //__intptr_t a; // flexible array member: array contiguous with the struct

 void R(int r) 
 if(*r)
 std::reverse(a, a+l); // *r && std::reverse() doesn't work because it returns void.

 if(d) // recurse if this isn't the bottom depth
 for(int i=0 ; i<l ; i++) // tree traversal
 ((m*)a[i])->R(r+1); // with the rest of the depth list
 

; // struct m

When we recurse, we pass r+1, so checking the current depth is always *r.

Reversing the current level before or after the tree traversal shouldn't matter. I didn't try to do it during.

I'm not sure that std::reverse is worth the byte count vs. a manual loop, especially if I can work in calling R() on each pointer exactly once somewhere inside that loop. But only if d!=0

edited Sep 5 at 8:26

answered Sep 5 at 7:53

Peter Cordes

1,912917

Whoa, impressive.
â€“Â AdÃ¡m
Sep 5 at 8:08

@AdÃ¡m: thanks, it golfed down surprisingly naturally and well after I wrote it. I'm curious to see what I can do in x86 machine code :P
â€“Â Peter Cordes
Sep 5 at 8:29

add a commentÂ |Â

up vote
1
down vote

(untested but I think correct. compiler asm output looks like what I expect. Will update if/when I find time to write a test harness that creates and prints this data structure.)

GNU C++ (portable) 148 bytes

#include<algorithm>
#include<cstdint>
struct mintptr_t d,l,a;void R(int*r)if(*r)std::reverse(a,a+l);for(int i=0;d&&i<l;((m*)a[i++])->R(r+1));;

GNU C++ (int=pointer and falls off a non-void function UB) 120 bytes

#include<algorithm>
struct mint d,l,a,R(int*r)if(*r)std::reverse(a,a+l);for(int i=0;d&&i<l;((m*)a[i++])->R(r+1));;

The R() reverse function is a member function because that saves declaring an arg, and saves a lot of p-> syntax to reference the struct members vs. the implicit this pointer.

The shorter version with just int also declares R() as returning int instead of void. These two bits of hackery are unrelated; this is just the "works on at least one implementation" version.

Ungolfed

The function arg, int r, is an array of 0 / non-zero integers that indicate whether a given depth should be swapped, starting with the outer-most level.

#include<algorithm> // for std::reverse
#include<cstdint> // for intptr_t. GNU C defines __intptr_t, so we could use that...

struct m
 __intptr_t d,l,a; // depth = 0 means values, >0 means pointers.
 // l = length
 //__intptr_t a; // flexible array member: array contiguous with the struct

 void R(int r) 
 if(*r)
 std::reverse(a, a+l); // *r && std::reverse() doesn't work because it returns void.

 if(d) // recurse if this isn't the bottom depth
 for(int i=0 ; i<l ; i++) // tree traversal
 ((m*)a[i])->R(r+1); // with the rest of the depth list
 

; // struct m

When we recurse, we pass r+1, so checking the current depth is always *r.

Reversing the current level before or after the tree traversal shouldn't matter. I didn't try to do it during.

I'm not sure that std::reverse is worth the byte count vs. a manual loop, especially if I can work in calling R() on each pointer exactly once somewhere inside that loop. But only if d!=0

edited Sep 5 at 8:26

answered Sep 5 at 7:53

Peter Cordes

1,912917

(untested but I think correct. compiler asm output looks like what I expect. Will update if/when I find time to write a test harness that creates and prints this data structure.)

GNU C++ (portable) 148 bytes

#include<algorithm>
#include<cstdint>
struct mintptr_t d,l,a;void R(int*r)if(*r)std::reverse(a,a+l);for(int i=0;d&&i<l;((m*)a[i++])->R(r+1));;

GNU C++ (int=pointer and falls off a non-void function UB) 120 bytes

#include<algorithm>
struct mint d,l,a,R(int*r)if(*r)std::reverse(a,a+l);for(int i=0;d&&i<l;((m*)a[i++])->R(r+1));;

The R() reverse function is a member function because that saves declaring an arg, and saves a lot of p-> syntax to reference the struct members vs. the implicit this pointer.

The shorter version with just int also declares R() as returning int instead of void. These two bits of hackery are unrelated; this is just the "works on at least one implementation" version.

Ungolfed

The function arg, int r, is an array of 0 / non-zero integers that indicate whether a given depth should be swapped, starting with the outer-most level.

#include<algorithm> // for std::reverse
#include<cstdint> // for intptr_t. GNU C defines __intptr_t, so we could use that...

struct m
 __intptr_t d,l,a; // depth = 0 means values, >0 means pointers.
 // l = length
 //__intptr_t a; // flexible array member: array contiguous with the struct

 void R(int r) 
 if(*r)
 std::reverse(a, a+l); // *r && std::reverse() doesn't work because it returns void.

 if(d) // recurse if this isn't the bottom depth
 for(int i=0 ; i<l ; i++) // tree traversal
 ((m*)a[i])->R(r+1); // with the rest of the depth list
 

; // struct m

When we recurse, we pass r+1, so checking the current depth is always *r.

Reversing the current level before or after the tree traversal shouldn't matter. I didn't try to do it during.

I'm not sure that std::reverse is worth the byte count vs. a manual loop, especially if I can work in calling R() on each pointer exactly once somewhere inside that loop. But only if d!=0

edited Sep 5 at 8:26

answered Sep 5 at 7:53

Peter Cordes

1,912917

edited Sep 5 at 8:26

answered Sep 5 at 7:53

Peter Cordes

1,912917

answered Sep 5 at 7:53

Peter Cordes

1,912917

answered Sep 5 at 7:53

Peter Cordes

1,912917

Whoa, impressive.
â€“Â AdÃ¡m
Sep 5 at 8:08

@AdÃ¡m: thanks, it golfed down surprisingly naturally and well after I wrote it. I'm curious to see what I can do in x86 machine code :P
â€“Â Peter Cordes
Sep 5 at 8:29

add a commentÂ |Â

Whoa, impressive.
â€“Â AdÃ¡m
Sep 5 at 8:08

@AdÃ¡m: thanks, it golfed down surprisingly naturally and well after I wrote it. I'm curious to see what I can do in x86 machine code :P
â€“Â Peter Cordes
Sep 5 at 8:29

Whoa, impressive.
â€“Â AdÃ¡m
Sep 5 at 8:08

@AdÃ¡m: thanks, it golfed down surprisingly naturally and well after I wrote it. I'm curious to see what I can do in x86 machine code :P
â€“Â Peter Cordes
Sep 5 at 8:29

add a commentÂ |Â

up vote
1
down vote

Mathematica, 7 bytes

Reverse

Function. Give it a nested list as the first argument, and the 1-based list of levels/dimensions to reverse as the second argument. Try it online!

^{^{Finally, another challenge where Mathematica has a builtin!}}

edited Sep 5 at 10:34

answered Sep 4 at 22:44

LegionMammal978

14.8k41752

layers to reverse? TIO link maybe?
â€“Â AdÃ¡m
Sep 5 at 1:52

@AdÃ¡m That is, the list of dimensions to reverse, generally referred to as levels in Mathematica.
â€“Â LegionMammal978
Sep 5 at 10:18

add a commentÂ |Â

up vote
1
down vote

Mathematica, 7 bytes

Reverse

Function. Give it a nested list as the first argument, and the 1-based list of levels/dimensions to reverse as the second argument. Try it online!

^{^{Finally, another challenge where Mathematica has a builtin!}}

edited Sep 5 at 10:34

answered Sep 4 at 22:44

LegionMammal978

14.8k41752

layers to reverse? TIO link maybe?
â€“Â AdÃ¡m
Sep 5 at 1:52

@AdÃ¡m That is, the list of dimensions to reverse, generally referred to as levels in Mathematica.
â€“Â LegionMammal978
Sep 5 at 10:18

add a commentÂ |Â

up vote
1
down vote

Mathematica, 7 bytes

Reverse

Function. Give it a nested list as the first argument, and the 1-based list of levels/dimensions to reverse as the second argument. Try it online!

^{^{Finally, another challenge where Mathematica has a builtin!}}

edited Sep 5 at 10:34

answered Sep 4 at 22:44

LegionMammal978

14.8k41752

Mathematica, 7 bytes

Reverse

Function. Give it a nested list as the first argument, and the 1-based list of levels/dimensions to reverse as the second argument. Try it online!

^{^{Finally, another challenge where Mathematica has a builtin!}}

edited Sep 5 at 10:34

answered Sep 4 at 22:44

LegionMammal978

14.8k41752

edited Sep 5 at 10:34

answered Sep 4 at 22:44

LegionMammal978

14.8k41752

answered Sep 4 at 22:44

LegionMammal978

14.8k41752

answered Sep 4 at 22:44

LegionMammal978

14.8k41752

layers to reverse? TIO link maybe?
â€“Â AdÃ¡m
Sep 5 at 1:52

@AdÃ¡m That is, the list of dimensions to reverse, generally referred to as levels in Mathematica.
â€“Â LegionMammal978
Sep 5 at 10:18

add a commentÂ |Â

layers to reverse? TIO link maybe?
â€“Â AdÃ¡m
Sep 5 at 1:52

@AdÃ¡m That is, the list of dimensions to reverse, generally referred to as levels in Mathematica.
â€“Â LegionMammal978
Sep 5 at 10:18

layers to reverse? TIO link maybe?
â€“Â AdÃ¡m
Sep 5 at 1:52

@AdÃ¡m That is, the list of dimensions to reverse, generally referred to as levels in Mathematica.
â€“Â LegionMammal978
Sep 5 at 10:18

add a commentÂ |Â

draft saved

draft discarded

draft saved

draft discarded

Post as a guest

Name

Category

Random preview

Multi-dimensional reversal

Walked-through example

Test cases

Walked-through example

Test cases

Walked-through example

Test cases

Walked-through example

Test cases

15 Answers 15

APL (Dyalog), 20 9 bytes

APL (Dyalog Unicode), 9 bytes

JavaScript (Node.js), 58 55 53 45 bytes

Commented

Python 2, 56 55 bytes

Jelly, 8 bytes

How it works

R, 80 78 77 bytes

Haskell, 120 119 bytes

05AB1E, 23 11 10 bytes

JavaScript (Node.js), 60 bytes

Pyth, 15 bytes

Japt, 15 14 bytes

Explanation

Clean, 122 112 bytes

Ruby, 54 bytes

GNU C++ (portable) 148 bytes

GNU C++ (int=pointer and falls off a non-void function UB) 120 bytes

Ungolfed

Mathematica, 7 bytes

Your Answer

Sign up or log in

Post as a guest

Post as a guest

15 Answers 15

15 Answers 15

APL (Dyalog), 20 9 bytes

APL (Dyalog), 20 9 bytes

APL (Dyalog), 20 9 bytes

APL (Dyalog), 20 9 bytes

APL (Dyalog Unicode), 9 bytes

APL (Dyalog Unicode), 9 bytes

APL (Dyalog Unicode), 9 bytes

APL (Dyalog Unicode), 9 bytes

JavaScript (Node.js), 58 55 53 45 bytes

Commented

JavaScript (Node.js), 58 55 53 45 bytes

Commented

JavaScript (Node.js), 58 55 53 45 bytes

Commented

JavaScript (Node.js), 58 55 53 45 bytes

Commented

Python 2, 56 55 bytes

Python 2, 56 55 bytes

Python 2, 56 55 bytes

Python 2, 56 55 bytes

Jelly, 8 bytes

How it works

Jelly, 8 bytes

How it works

Jelly, 8 bytes

How it works

Jelly, 8 bytes

How it works

R, 80 78 77 bytes

R, 80 78 77 bytes

R, 80 78 77 bytes

R, 80 78 77 bytes

Haskell, 120 119 bytes

Haskell, 120 119 bytes

Haskell, 120 119 bytes

Haskell, 120 119 bytes

05AB1E, 23 11 10 bytes

05AB1E, 23 11 10 bytes

05AB1E, 23 11 10 bytes

05AB1E, 23 11 10 bytes

JavaScript (Node.js), 60 bytes

JavaScript (Node.js), 60 bytes

15 Answers
15

15 Answers
15

15 Answers
15